Research Edition · Systematic Literature Map

Cognitive Offloading & AI

Every discoverable arXiv and scholarly research paper in the strict corpus—compressed to one insight and one conclusion, with access and review status kept visible.
Through August 13, 2026  ·  262 papers  ·  57 arXiv  ·  104 full text  ·  158 abstract only
✎  Education (147)
2026-05-30 · Zenodo (CERN European Organization for Nuclear Research) · Conceptual
莉莉娜大主教的恶意
  • Insight. Through integrating empirical data from 2023—2026, this paper provides multi-domain evidence of AI deskilling: the "code blindness" phenomenon in software engineering, the loss of cognitive sovereignty in medical diagnosis, the hallucination crisis in legal practice, the atrophy of critical thinking in education, and the floor-zeroing pattern of autonomous driving systems.
  • Conclusion. The study argues that AI deskilling is not merely traditional technological substitution, but an irreversible surrender of cognitive sovereignty.
EducationConceptualabstract onlypreprint / unreviewed
2026-02-17 · arXiv (Cornell University) · Conceptual
C. K. Y Chan
  • Insight. Using LLMs to bring evidence-based feedback into the classroom: AI-generated feedback increases secondary students’ text revision, motivation, and positive emotions.
  • Conclusion. Future research should move from conceptual description toward design and intervention studies that support learners in maintaining epistemic agency when working with AI.
EducationConceptualfull textpreprint / unreviewed
2026-05-30 · BMC Psychology · Empirical
Wenlong Wang, Yuhang Wu, Jie Fang, YANG Chong + 1 more
  • Insight. These findings suggest that excessive reliance on AI as a stress-coping strategy may weaken students' sense of competence and increase their vulnerability to academic burnout and anxiety.
  • Conclusion. Educational interventions should therefore promote reflective and autonomy-supportive AI use, positioning AI as a learning scaffold rather than a substitute for students' own cognitive engagement.
EducationEmpiricalabstract onlypublisher record
2026-06-24 · Journal of Intelligence · Empirical
S Si, Yong Qi, Jingming Xu, Xinyu Qi
  • Insight. Because all constructs were measured through self-report, the findings are interpreted as reflecting students’ perceptions of their cognitive behaviors and abilities; the methodological implications of this design are discussed in detail.
  • Conclusion. Educational interventions should prioritize metacognitive training over technical skill development to ensure that cognitive offloading enhances rather than undermines critical thinking.
EducationEmpiricalabstract onlypublisher record
2026-07-30 · arXiv · Research article
Amir Moradifam
  • Insight. The challenge concerns not only academic integrity but also the risks of cognitive offloading and the illusion of explanatory depth.
  • Conclusion. Mathematics is often valued for its role as the language of science, but it is equally important as a human practice that teaches us how to think and cultivates creativity, problem-solving skills, attention, and judgment.
EducationResearch articlefull textpreprint / unreviewed
2026-06-22 · International Journal of Emerging and Disruptive Innovation in Education VISIONARIUM · Research article
Tina R. Austin
  • Insight. When students are rewarded for polished outputs, cognitive offloading to AI tools becomes rational, and traditional snapshot assessments (single-moment evaluations of task completion) produce false signals about whether durable learning has occurred.
  • Conclusion. Telling students to “reflect on their own thinking” has become insufficient in AI-mediated learning environments.
EducationResearch articlefull textpublisher record
2026-01-01 · SSRN Electronic Journal · Research article
Jack Jassan
  • Insight. Two competing hypotheses are evaluated: (1) the Regression Hypothesis, suggesting diminished internal structural rehearsal, and (2) the Redistribution Hypothesis, proposing cognitive reallocation rather than decline.
  • Conclusion. By integrating literature on cognitive offloading, neuroplasticity, executive function, and the extended mind, this paper argues that generative AI represents an unprecedented expansion of cognitive externalization with long-term implications for education, intellectual development, and inequality.
EducationResearch articleabstract onlypreprint / unreviewed
2026-01-27 · Assessment & Evaluation in Higher Education · Research article
Anne-Marie Chase, K. M. Galvin
  • Insight. In doing so, it addresses the risks of cognitive offloading and reinforces the validity of assessment practices across diverse disciplinary contexts.
  • Conclusion. Ultimately, this paper argues that the future of assessment in the GenAI era must be shaped not by compliance alone, but by care, clarity, and critical engagement.
EducationResearch articlefull textpublisher record
2026-05-28 · Pacific Journal of Technology Enhanced Learning · Empirical
Anne-Marie Chase, K. M. Galvin
  • Insight. Evidence of outcomes and effectiveness Preliminary findings indicate that participants experience improved clarity about what constitutes evidence of learning, increased confidence in articulating permissible GenAI use to students and reduced ambiguity when applying AI assessment scales.
  • Conclusion. However, recent scholarship has argued that validity-focused, design-led reform is required if assessment is to remain meaningful in AI-rich environments (Corbin et al., 2025; Dawson et al., 2024).
EducationEmpiricalfull textpublisher record
2026-01-30 · arXiv · Empirical
Rudrajit Choudhuri, Christopher Sanchez, Margaret Burnett, Anita Sarma
  • Insight. In survey data from 299 STEM students at five North American universities, routine GenAI use and greater trust were associated with less reflection, less need for understanding, and less critical thinking in coursework.
  • Conclusion. Technophilic motivation, risk tolerance, and computer self-efficacy did not protect students; those traits were associated with greater vulnerability to disengagement, while prior AI or academic experience offered no clear buffer.
EducationEmpiricalfull textpreprint / unreviewed
2026-01-02 · arXiv (Cornell University) · Research article
Aizierjiang Aiersilan
  • Insight. By quantifying skill retention, error detection, and conceptual understanding, VCP addresses the critical need for evidence-based assessment in the era of Vibe Coding.
  • Conclusion. The Vibe-Check Protocol offers a systematic framework for evaluating the educational impact of AI-assisted programming.
EducationResearch articlefull textpreprint / unreviewed
2026-04-28 · SISTEMASI · Research article
RZ Abdul Aziz, Anton Setiawan
  • Insight. The findings indicate that creativity, AI, and critical thinking positively and significantly influence cognitive offloading.
  • Conclusion. This study investigates the role of cognitive offloading in mediating the effects of creativity, artificial intelligence (AI), and critical thinking on Higher Order Thinking Skills (HOTS).
EducationResearch articleabstract onlypublisher record
2026-01-01 · SSRN Electronic Journal · Research article
Dr. Joji Valli
  • Insight. This article examines the emergent psychological paradigm of "AI-Native" cognition—a cognitive framework developing in individuals whose formative psychological and intellectual development occurs alongside advanced artificial intelligence systems.
  • Conclusion. We argue that AI is not merely a tool but a constitutive element of a new hybrid cognitive ecology.
EducationResearch articleabstract onlypublisher record
2026-07-18 · Zenodo (CERN European Organization for Nuclear Research) · Conceptual
SK Haque Mobassir Imtiyazul Haque
  • Insight. This paper identifies the integration of artificial intelligence (AI) and continuous automation across educational, vehicular, and consumer mobile ecosystems as a permanent, irreversible shift in human cognitive habits.
  • Conclusion. As spatial navigation, sensory filtering, and critical analytical tasks are systematically transferred to algorithmic processing layers, traditional metrics of human intelligence undergo structural changes.
EducationConceptualabstract onlypublisher record
2026-04-16 · arXiv (Cornell University) · Research article
Hyunwoo Kim, Harin Yu, Hanau Yi
  • Insight. This paper introduces the LLM fallacy, a cognitive attribution error in which individuals misinterpret LLM-assisted outputs as evidence of their own independent competence, producing a systematic divergence between perceived and actual capability.
  • Conclusion. We argue that the opacity, fluency, and low-friction interaction patterns of LLMs obscure the boundary between human and machine contribution, leading users to infer competence from outputs rather than from the processes that generate them.
EducationResearch articlefull textpreprint / unreviewed
2026-04-10 · Zenodo (CERN European Organization for Nuclear Research) · Empirical
PYELSHAK YUSUF*, Natherniel Pam, Josiah Nanpon, Bala Yakubu + 2 more
  • Insight. As generative AI becomes increasingly prevalent in educational settings, these findings underscore the urgent need for pedagogical frameworks that harness the benefits of AI while safeguarding the development of students' metacognitive autonomy and capacity for lifelong learning.
  • Conclusion. The rapid integration of generative artificial intelligence (AI) tools in higher education has prompted significant scholarly debate regarding their pedagogical implications.
EducationEmpiricalabstract onlypublisher record
2026-05-15 · Minnesota Journal of Business Law and Entrepreneurship · Empirical
Priyam Kukreja, P. Soumya, Subhash Kutty, Prof. Ritu Lund
  • Insight. However, this convenience has given rise to a new phenomenon called “market–ready atrophy” which shows the widening gap between students' perceived competence and their actual ability to solve problems without the help of algorithms.
  • Conclusion. There is a prominent shift in how management education students approach analytical tasks of solving complex business problems, market research interpretations, evaluating operational efficiency or case study analysis due to the rapid integration of generative Artificial Intelligence (Gen Ai) in higher education.
EducationEmpiricalabstract onlypublisher record
2026-06-30 · Asian Journal of Current Research · Research article
Hammed Adekunle Abdulazeez, Kehinde Daniel Obidele, Naomi Ama Kedador, Esther Uyoyooghene Olokede
  • Insight. There is evidence of weakened memory consolidation, increased automation bias, shifts in how people judge their own competence, and patterns resembling digital dependency, alongside genuine gains in productivity and reduced cognitive strain under the right conditions.
  • Conclusion. The discussion closes with implications for education, professional training, clinical psychology, and AI governance, and identifies where future research is most needed.
EducationResearch articleabstract onlypublisher record
2026-06-23 · Preprints.org · Conceptual
Dimitar Trajanov, Ljupčo Kocarev
  • Insight. We identify five early, individually preliminary signals consistent with that hypothesis: (1) reduced neural engagement during AI-assisted writing; (2) cognitive offloading and lower critical-thinking scores among adults who delegate reasoning to chatbots; (3) AI dependence associated with weaker reasoning among students; (4) deskilling in knowledge work, accelerated as passive assistants give way to agents that perform whole tasks; and (5) the displacement of the desirable difficulties known to consolidate learning.
  • Conclusion. Reading the physical case as a worked example, we argue that the same three-step mechanism that drove bodily decline is now plausibly acting on cognition: substitution of effort, redesign of the surrounding environment, and a rebound in which a cheaper resource is consumed in greater quantity (the Jevons paradox).
EducationConceptualabstract onlypreprint / unreviewed
2026-05-09 · Open MIND · Research article
Dr. David Bull
  • Insight. Drawing on cognitive psychology, human–computer interaction, and philosophy, the model identifies key mechanisms, including cognitive offloading, automation bias, and cognitive surrender, through which AI reshapes thinking.
  • Conclusion. The framework further argues that AI operates as a dual system, simultaneously enabling knowledge expansion while introducing epistemic risks such as illusion of understanding and diminished intellectual autonomy.
EducationResearch articleabstract onlypublisher record
2026-01-01 · SSRN Electronic Journal · Empirical
Inara Scott
  • Insight. Drawing on cognitive offloading theory, desirable difficulties research, and emerging experimental evidence on GenAI and learning, the framework predicts that AI use structured around executor roles, where judgment and critical thinking reside primarily with the AI tool, poses the highest risk of cognitive offloading and reduced critical thinking, while evidence and opponent roles (which structurally require the learner to evaluate, defend, and revise) are most likely to support durable learning.
  • Conclusion. The paper presents role definitions, classification rules, risk mechanisms, design principles, and practical assignment examples.
EducationEmpiricalabstract onlypublisher record
2026-06-09 · Zenodo (CERN European Organization for Nuclear Research) · Conceptual
Nathaniel Lukas Jonathan Berky
  • Insight. Keywords: generative AI · AI literacy · cognitive offloading · developmental neuroscience · educational technology · 4D framework · Constitutional AI · AI ethics in education · adolescent executive function · OECD AILit framework · PISA 2029 ·.
  • Conclusion. The paper concludes with operational guidance for institutions across the 2026–2030 horizon, a limitations section, and an agenda for future empirical research.
EducationConceptualabstract onlypublisher record
2026-05-08 · International Journal of English Linguistics · Empirical
Smaoui Anouar, Bensalem Hanen
  • Insight. The results indicate that GenAI supports SRIL not through availability alone, but through instructional mediation that transforms AI use from procedural reliance into regulated, reflective, and ethically grounded academic writing practice.
  • Conclusion. Scaffolded generative AI (GenAI) in higher education academic writing has attracted growing scholarly attention, yet empirical evidence remains limited regarding its effects on self-regulated intercultural learning (SRIL).
EducationEmpiricalfull textpublisher record
2026-04-27 · IEEE Global Engineering Education Conference · Empirical
Rafael Emilio Davalos-Villarreal, Humberto Genaro Cisneros Salinas, Jorge Adrian De La Garza Flores
  • Insight. A case study of a final IoT project and a survey conducted with Computer Science students were used to analyze how a scaffolded CBL framework guides responsible AI use.Findings indicate that when AI is utilized responsibly as a tool to aid for obstacles beyond the syllabus, students engage more deeply, develop stronger technical skills, and concentrate on solution design without limiting creativity due to technical gaps.
  • Conclusion. AI tools can improve efficiency and enrich the learning experience, but they risk encouraging student over-reliance, which can obstruct the development of critical thinking and problem-solving skills.
EducationEmpiricalabstract onlypublisher record
2026-06-14 · arXiv (Cornell University) · Conceptual
Fendi Tsim, Alina Gutoreva
  • Insight. Chatgpt as a cognitive crutch: Evidence from a randomized controlled trial on knowledge retention.
  • Conclusion. Cognitive research: principles and implications, 9 (1), 20. https://doi.org/10.1186/s41235-024-00547-9 Baek, S.-G., & Kim, K.
EducationConceptualfull textpreprint / unreviewed
2026-02-23 · arXiv (Cornell University) · Research article
Olga Viberg, Mutlu Cukurova, Rene F. Kizilcec, Simon Buckingham Shum + 11 more
  • Insight. We explore practical dilemmas for protecting and promoting agency, focusing on normative constraints, transparency, and cognitive offloading, and highlight key tensions and implications to inform ethical and effective AI integration in education.
  • Conclusion. Human agency is crucial in education and increasingly challenged by the use of generative AI.
EducationResearch articlefull textpreprint / unreviewed
2026-05-20 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Flavia Mandarino2 Antonella Balena1*
  • Insight. Results also suggest that combining textual prompts with digital resources enhances specificity and supports multimodal learning.
  • Conclusion. The study suggests that prompt engineering may represent an emerging pedagogical competence for inclusive language teaching and calls for further research on the long-term cognitive and educational implications of Artificial Intelligence-supported instruction.
EducationResearch articleabstract onlypublisher record
2026-06-09 · arXiv (Cornell University) · Research article
Oleksandra Poquet, Mani Shankar Nanduri, Maria Ximena Salinas Loyer, Matthias Stadler + 2 more
  • Insight. Volume-based profiling (k=4) differentiated learners primarily by prior knowledge, with volume negatively associated with essay authorship.
  • Conclusion. Combining volume and content of offloading could improve future analyses on how LLM use redistributes cognitive activity and its effects on learners.
EducationResearch articlefull textpreprint / unreviewed
2026-06-30 · Psychosomatic Medicine and General Practice · Empirical
А. В. Чорномидз, Illia Ilchuk
  • Insight. Keywords: artificial intelligence, generative models, problematic use, cognitive offloading, dissimulation, medical students. УДК: 616.89-008.44:004.8
  • Conclusion. A questionnaire including a targeted block of statements regarding the nature of interaction with algorithms was utilized to assess daily practices and risks.
EducationEmpiricalfull textpublisher record
2026-04-12 · arXiv (Cornell University) · Empirical
Neha Rani, Erta Cenko, Laura Melissa Cruz Castro
  • Insight. Our results are based on empirical evidence from en- gineering students’ perceptions.
  • Conclusion. Through our investigation, we aim to understand the dynamically changing space of AI integration in computing and its impacts on the importance of cognitive skills and, therefore, the reshaping of the future computing workforce.
EducationEmpiricalfull textpreprint / unreviewed
2026-03-24 · International Journal of Educational Technology in Higher Education · Empirical
Shaofeng Wang, Hao Zhang
  • Insight. The findings reveal a counterintuitive paradox: pedagogical partnerships simultaneously activate cognitive vigilance (critical evaluation) and cognitive offloading (strategic delegation).
  • Conclusion. Furthermore, efficiency orientation acts not as a barrier but as an amplifier, strengthening both vigilance and offloading behaviors through mechanisms of pragmatic risk management.
EducationEmpiricalfull textpublisher record
2026-07-06 · Edelweiss Applied Science and Technology · Empirical
Chenyi Liao, Hongchun Du, Jian Yao
  • Insight. Results show prevalent multi-GenAI adoption among students (only 14.1% rarely use such tools), with ChatGPT, DeepSeek, Kimi, Gemini, and Doubao mainly used for grammar, translation, and vocabulary learning.
  • Conclusion. Generative AI (GenAI) has proliferated in language education, yet empirical research on its real-world use remains scarce, especially beyond mainstream EFL settings.
EducationEmpiricalfull textpublisher record
2026-04-26 · Center for Open Science · Research article
Liudmila Piatnitckaia, Valentin Corneloup, Babette Bühler, Nađa Terzimehić + 2 more
  • Insight. Results showed that “Heavy offloading” clustered in “Create” dialogues (e.g., drafting arguments or code) and appeared more often among students who receive high grades; lower-tier students favored light or no offloading.
  • Conclusion. Generative AI is increasingly woven into students’ everyday academic practices, yet the extent of cognitive offloading, and at what cognitive levels remain insufficiently understood.
EducationResearch articleabstract onlypreprint / unreviewed
2026-03-29 · Inverge Journal of Social Sciences · Research article
Muhammad Faisal Ishaque, Amana Ishaaq, Maleeha Nazim
  • Insight. These results showed that AI-tool users have improved critical thinking scores by a significant margin, 51.5 (SD = 6.3) to 68.0 (SD = 5.0) (t = 8.72, p < 0.05), and the control group did not significantly improve in critical thinking scores.
  • Conclusion. Adaptive Marketing Systems and Consumer Feedback Loops: Implications for Market Development in Emerging Economies.
EducationResearch articleabstract onlypublisher record
2026-01-01 · International Research Journal of Multidisciplinary Scope · Empirical
Vengalarao Pachava, Olusiji Adebola Lasekan, Margot Teresa Godoy Pena, Kotigari Reddi Swaroop + 1 more
  • Insight. While LLMs enhance accessibility and efficiency, findings indicate that excessive reliance encourages cognitive offloading, weakens critical thinking, and reshapes students’ learning practices toward surface-level engagement.
  • Conclusion. By foregrounding students’ lived experiences, this research contributes to the growing discourse on digital transformation in higher education and underscores the need to align LLM integration with responsible learning practices that strengthen intellectual autonomy and academic integrity.
EducationEmpiricalfull textpublisher record
2026-02-02 · ACADEMIA International Journal for Social Sciences · Empirical
Basit Aziz, Syeda Laraib Zehra, Muhammad Hassan Shafiq
  • Insight. The mixed-methods study reports that heavier ChatGPT dependence was associated with lower critical-thinking engagement, weaker confidence in unaided writing, and greater psychological dependence and anxiety.
  • Conclusion. It recommends explicit institutional rules and learning designs that preserve independent writing and reasoning; the reported associations do not by themselves establish that ChatGPT caused the differences.
EducationEmpiricalfull textpublisher record
2026-05-09 · Development in Learning Organizations An International Journal · Research article
Michael Shick, Janice A. Black
  • Insight. Findings Accelerating AI adoption brings enhanced rationality closer to realization yet risks over-reliance on decision-support and diminished independent reasoning.
  • Conclusion. Practical implications Management educators should integrate guided AI applications in assignments.
EducationResearch articleabstract onlypublisher record
2026-04-29 · Figshare · Empirical
KEYI LIU
  • Insight. Results showed that generative AI-assisted lesson planning was positively associated with productive cognitive offloading, but this pathway did not translate into higher pedagogical creativity.
  • Conclusion. Generative artificial intelligence (AI) is increasingly visible in teachers’ lesson planning, yet less is known about the cognitive mechanisms through which AI-assisted planning is linked to teachers’ pedagogical creativity.
EducationEmpiricalabstract onlypublisher record
2026-07-16 · Frontiers in Psychology · Empirical
Qiuhan Zhu, Xiangnan Li, Yiang Dong, Pengcheng Chang + 1 more
  • Insight. These findings highlight the manner of AI engagement—not merely its frequency—as a key factor in understanding its associations with perceived cognitive functioning, and point to practical strategies for educators, learners, and AI tool designers seeking to harness AI without undermining cognitive autonomy.
  • Conclusion. Generative AI tools such as ChatGPT now perform core cognitive operations—reasoning, synthesis, evaluation, and creative generation—on users' behalf, raising urgent questions for educational psychology about how AI use relates to cognitive development.
EducationEmpiricalfull textpublisher record
2026-08-11 · Review of Artificial Intelligence in Education · Empirical
Abd Wahab, Thomas Asabere, Issah Chakurah
  • Insight. Uncalibrated trust was positively associated with cognitive offloading (β = .319), which in turn reduced learning outcomes (β = -.216).
  • Conclusion. The findings support a cognitively informed account of human-AI interaction and underscore the need for AI literacy interventions that cultivate critical engagement rather than passive dependence, particularly within Global South higher education contexts.
EducationEmpiricalabstract onlypublisher record
2026-01-01 · Open MIND · Research article
Chun Yuen Fong
  • Insight. First, it aims to provide empirical evidence on whether and how sustained AI use is associated with shifts in core aspects of self-regulation.
  • Conclusion. Second, it seeks to contribute to a broader understanding of cognitive development in AI-assisted environments, with implications for educational practices and policy.
EducationResearch articleabstract onlypreprint / unreviewed
2026-06-16 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Geiger, Lukas
  • Insight. Abstract (English) Background: The MIT study "Your Brain on ChatGPT" (Kos'myna et al., 2025) reported decreased neural activity when participants used ChatGPT for essay writing, interpreting this as "cognitive debt." Objective: This critical analysis examines whether the study's methodology and conclusions are warranted, contextualizing the findings within the broader evidence on cognitive effects of LLM usage.
  • Conclusion. A new bilingual evidence ledger was added after the 2x2 taxonomy to link central inference claims to their strongest evidence anchors, current status, remaining gaps, and design implications.
EducationResearch articleabstract onlypreprint / unreviewed
2026-06-17 · Behavioral Sciences · Empirical
Wenqiang Fan, Lu Cheng, Yue Wang, Qi Zhao + 1 more
  • Insight. The results show that cognitive relief exerts a complementary mediating effect between AI use and attitudes, whereas cognitive offloading functions as a competitive mediator.
  • Conclusion. Moreover, educators and institutions should leverage AI to provide cognitive relief for higher-order learning activities, thereby enhancing the engagement, motivation, and deeper learning processes of students, while simultaneously implementing reflective and critical thinking practices to guard against the risks of cognitive offloading.
EducationEmpiricalabstract onlypublisher record
2026-07-11 · Indonesian Journal of Science and Mathematics Education · Empirical
Ardian Asyhari, R. Diani, Nor Farawahidah Abdul Rahman
  • Insight. HITL-AI also showed higher argumentation quality, metacognitive monitoring, and germane load, alongside lower AI overreliance and cognitive offloading than AI-only learning.
  • Conclusion. AI integration in physics learning should be structured through instructor-mediated argumentation protocols to preserve epistemic agency and strengthen higher-order thinking.
EducationEmpiricalabstract onlypublisher record
2026-01-01 · Open MIND · Empirical
Chaeyun Lim, Diego Gómez-Zará, Dashun Wang
  • Insight. Evidence shows that students tend to skip engaging with task material and rely on AI to supply answers, which hurts their performance once AI access is removed and they must work independently (Bastani et al., 2025).
  • Conclusion. This research project therefore aims to address the following research questions: Overarching RQ1: Do people skip pre-task knowledge acquisition and rely on AI during execution when a task requires learning new knowledge and skills before beginning?
EducationEmpiricalabstract onlypreprint / unreviewed
2026-01-01 · SSRN Electronic Journal · Research article
Ahmad Alamgir Khan
  • Insight. These constructs describe how fluent AI-generated outputs influence perceived correctness, reduce independent verification, and increase reliance on machine-generated reasoning over time.
  • Conclusion. A Domain Sensitivity Coefficient (DSC) is introduced to capture how these effects vary across domains with different risk levels and verification costs.
EducationResearch articleabstract onlypreprint / unreviewed
2026-07-09 · arXiv (Cornell University) · Research article
Shayla Sharmin, Mohammad Al-Ratrout, Mohammad Fahim Abrar, Roghayeh Leila Barmaki
  • Insight. These findings reveal a structural tension in self-directed AI learning: content that prioritizes quick outputs reaches far more learners than content that promotes deep engagement.
  • Conclusion. We examine educational YouTube videos through multimodal metadata, such as transcripts, titles, thumbnails, and viewer comments, to investigate how ChatGPT is framed across creator groups and how those framings relate to audience response and platform reach.
EducationResearch articlefull textpreprint / unreviewed
2026-02-14 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Zeynep Jaggi
  • Insight. The results show that when AI feedback is embedded within a teacher-structured framework that demands explanation and accountability, learners engage more deeply with their own revision process.
  • Conclusion. The thesis concludes with practical implications for classroom use: prompts should invite reasoning, not simple correction; AI feedback should remain dialogic and reflective; and writing tasks should align with Lehrplan 21 (curriculum) competencies on strategy use and self-regulation.
EducationResearch articleabstract onlypublisher record
2026-05-18 · arXiv (Cornell University) · Conceptual
Shan Li, Juan Zheng
  • Insight. It draws on established scholarship in distributed cognition, extended mind, human-AI collaboration, AI literacy, cognitive offloading, and metacognition research.
  • Conclusion. Without an adequate theoretical framework, educators, designers, and policymakers lack the conceptual tools to make informed decisions about how generative AI should be integrated into learning.
EducationConceptualfull textpreprint / unreviewed
2026-05-27 · INPLASY - International Platform of Registered Systematic Review and Meta-analysis Protocols · Review
Guanhua Wang, Wenna Wang, Daozhou Yang, Jifan Ren
  • Insight. The protocol organizes a scoping review around two coexisting patterns: GenAI can support self-regulation and participation, or erode agency through dependence, uncritical uptake, and delegated judgment.
  • Conclusion. Its review question is whether learners still plan, monitor, verify, revise, and accept responsibility for AI-supported work; as a protocol, it should not be read as reporting completed review findings.
EducationReviewabstract onlypublisher record
2026-07-08 · Behavioral Sciences · Review
Guanhua Wang, W Wang, Daozhou Yang, Jifan Ren
  • Insight. Overall, the findings suggest that the educational value of GenAI depends less on the technology itself than on how it is pedagogically embedded, with augmentation-oriented and scaffolded use being more supportive of learner agency than replacement-oriented use.
  • Conclusion. These patterns should be interpreted as a configurative synthesis of heterogeneous evidence rather than as causal effect estimates.
EducationReviewabstract onlypublisher record
2026-06-11 · Humanities and Social Sciences Communications · Empirical
Zarqa Farooq Hashmi, Javed Iqbal, Muhammad Zaheer Asghar, Muhammad Naseem Abid
  • Insight. These findings support our research model, suggesting that TSE and CO play important mediating roles in the relationship between GenAITU and SRL.
  • Conclusion. Therefore, the practical implications of this research are most salient for educators and institutions designing blended curricula.
EducationEmpiricalfull textpublisher record
2026-06-07 · Interactive Learning Environments · Empirical
Hasan Zahid, Muhammad Danish, Sabir Ali, Chunhui Huo + 1 more
  • Insight. The findings reveal that six gratifications, perceived usefulness, perceived ease of use, personalization, trust, emotional attachment, and anthropomorphism significantly predict GAIC addiction, which in turn leads to increased cognitive offloading.
  • Conclusion. While prior studies have focused on adoption and continuance, limited attention has been given to the potential risks of overreliance.
EducationEmpiricalabstract onlypublisher record
2026-01-01 · OSF Preprints (OSF Preprints) · Review
Eyüp Yurt
  • Insight. Findings: Seventy-three percent of studies (n = 32) reported CL reduction, 14% (n = 6) reported mixed effects, and 7% (n = 3) documented CL increases.
  • Conclusion. Background: Generative artificial intelligence (GenAI) tools have rapidly pervaded higher education, yet the cognitive mechanisms through which they affect learning remain poorly understood.
EducationReviewabstract onlypreprint / unreviewed
2026-05-26 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Mushahid Khuhro
  • Insight. The findings suggest that excessive reliance on generative AI may reduce cognitive struggle, reflective inquiry, and independent reasoning by encouraging passive knowledge consumption and cognitive delegation.
  • Conclusion. The paper argues that generative AI should not be viewed merely as a productivity tool, but as a cognitive environment capable of reshaping how students construct, externalize, and engage with knowledge.
EducationResearch articleabstract onlypublisher record
2026-07-22 · Open MIND · Review
Nitya Verma, Harshith B Nair, Ashutosh Mohanty
  • Insight. The increasing integration of generative AI tools into everyday learning and research tasks has renewed interest in the psychological phenomenon of cognitive offloading — the tendency to use external tools to reduce internal cognitive demand.
  • Conclusion. Findings inform practical implications for educational design, teacher training, and assessment as AI tools become embedded in classroom preparation work, and surface open questions for larger-scale empirical research.
EducationReviewabstract onlypreprint / unreviewed
2026-05-10 · European Journal of Applied Science Engineering and Technology · Research article
Tran Minh Duc
  • Insight. IMPLICATIONS The results of this conceptual analysis provide several recommendations for universities, lecturers, and policymakers seeking to respond positively to the emergence of generative AI in higher education.
  • Conclusion. Addressing the cognitive impact of AI in education is therefore imperative for preserving the intellectual purpose of higher education in the future.
EducationResearch articlefull textpublisher record
2026-03-27 · International Conference on Educational and Information Technology · Empirical
Shu Chen, Fenghua Liu
  • Insight. These findings suggest that in GenAI-assisted environments, mandatory prompt structuring may serve as an external regulatory script.
  • Conclusion. Based on these empirical observations, the study proposes design principles for preserving germane cognitive load through gradual scaffold fading and the integration of process-oriented assessments in future human-GenAI collaborative programming interfaces.
EducationEmpiricalabstract onlypublisher record
2026-06-29 · Frontiers in Psychology · Empirical
Li Gao, Yanli Sun, Safeer Ullah Khan
  • Insight. The qualitative findings support and explain these results, highlighting the behavioral mechanisms underlying AI-supported learning.
  • Conclusion. Overall, the study demonstrates that the impact of AI on learning is dual in nature and depends on how students engage with the technology, offering important implications for theory and educational practice.
EducationEmpiricalfull textpublisher record
2026-06-01 · Interactive Learning Environments · Empirical
Yu Liu, Min Jou, Yungwei Hao
  • Insight. Findings indicate that greater AI use was associated with stronger cognitive offloading, which was further associated with greater epistemic trust transfer toward AI and lower learning accountability.
  • Conclusion. Reduced learning accountability was also associated with lower learning motivation and greater academic integrity risk.
EducationEmpiricalabstract onlypublisher record
2026-02-18 · 2026 6th International Conference on Advanced Research in Computing (ICARC) · Research article
N.M.A.H. Gunathilake, Amindya Gamage, Dilki Rajapakshe, M. Jayasooriya + 2 more
  • Insight. The findings reveal significant direct, indirect, and moderating relationships among key variables, suggesting that the use of AI-assisted learning tools influences students' learning regulation, reliance on external support, and development of higher-order thinking.
  • Conclusion. The rapid adoption of AI-assisted learning tools in higher education has completely transformed the undergraduate study system, but empirical evidence on their impact on deep cognition and learning processes is limited.
EducationResearch articleabstract onlypublisher record
2026-03-20 · arXiv (Cornell University) · Research article
Yonas Kassa, James Bonacci, Ping Wang
  • Insight. These challenges stem from concerns that LLMs undermine academic assessment by enabling bypassing of critical thinking, leading to increased cognitive offloading.
  • Conclusion. To support future research and development in this domain initial codes with a readily deployable 'AI-Sinkhole' blockist is available on https://github.com/AIMLEdu/ai-sinkhole.
EducationResearch articlefull textpreprint / unreviewed
2026-01-16 · arXiv (Cornell University) · Research article
Chan-Jin Chung
  • Insight. These preliminary results suggest that allowing GenAI for programming assignments does not diminish students' mastery of course concepts when learning is verified through targeted, assignment-driven quizzes.
  • Conclusion. Although limited by a small sample size, this study provides preliminary evidence that the risks of cognitive offloading can be mitigated by allowing AI-assisted programming practice while verifying understanding through assignment-driven, AI-free quizzes.
EducationResearch articlefull textpreprint / unreviewed
2026-07-20 · International Journal of Technology in Education · Empirical
J. Alamri, Waleed Salim Alhalafawy
  • Insight. The findings suggest that GenAI is more effective at supporting retention when used to scaffold the active reasoning of learners as opposed to substituting for their cognitive processing.
  • Conclusion. Even though the effects of Generative AI (GenAI) on learning retention are still not adequately understood, its use in higher education is increasing.
EducationEmpiricalabstract onlypublisher record
2026-03-31 · International Journal of Education Psychology and Counseling · Review
Nurkaliza Khalid, Rafiza Kasbun, Nurul Jannah Abd Rahman
  • Insight. First, it introduces a three-lens framework: Efficiency, Engagement, and Cognitive Offloading which provides a structured way to synthesize fragmented findings across diverse methodologies, disciplines, and educational levels.
  • Conclusion. Future research should prioritize experimental and longitudinal designs that can track sustained cognitive consequences across multiple time points.
EducationReviewfull textpublisher record
2026-02-03 · Frontiers in Medicine · Research article
Shereen El Tarhouny, Amira Farghaly
  • Insight. The hippocampus, crucial for memory formation and retrieval, shows reduced involvement, resulting in weaker encoding and recall of clinical information.
  • Conclusion. AI should therefore be adopted only when it enhances outcomes without compromising the relational dimensions of care.
EducationResearch articlefull textpublisher record
2026-01-01 · Inquiry Critical Thinking Across the Disciplines · Research article
Amanda Hiner
  • Insight. Recent research has revealed that over-reliance on GenAI is associated with both cognitive offloading and diminished critical thinking skills and traits, including intellectual autonomy and independent problem-solving skills.
  • Conclusion. Addressing the challenges and risks of GenAI use within educational contexts will require educators to modify their approaches to teaching digital literacy, sharpen students’ awareness of the cognitive implications of GenAI use, and encourage students to make context- and purpose-driven decisions about when and how to use AI technologies.
EducationResearch articleabstract onlypublisher record
2026-03-18 · arXiv.org · Empirical
M Murshidul Bari, Akif Islam, Mohd Ruhul Ameen, Abu Saleh Musa Miah + 1 more
  • Insight. The findings reveal a mixed pattern: participants largely viewed AI as a tool for speed, convenience, and learning support, yet many also reported reduced patience for sustained effort.
  • Conclusion. The paper therefore argues that effective human-AI collaboration should support reflection, verification, and sustained cognitive effort rather than substitute for them.
EducationEmpiricalfull textpreprint / unreviewed
2026-05-06 · arXiv (Cornell University) · Empirical
Tosin Adewumi, Marcus Liwicki, Foteini Simistira Liwicki, Lama Alkhaled + 2 more
  • Insight. The results reveal that (1) the students' self-written counterarguments to AI-generated content contains logic, among other things, which is a key component of critical thinking, and (2) GenAI can be successfully used at scale to assess students' written work, based on clear rubrics, and these assessments generally align with human assessments as shown with Gwets AC2 inter-rater reliability values of 0.33 for all the models except one.
  • Conclusion. This is especially as risks of cheating and cognitive offloading exist with the use of GenAI.
EducationEmpiricalfull textpreprint / unreviewed
2026-04-03 · arXiv (Cornell University) · Research article
Hanyu Su, Huilin Zhang, Shihui Feng
  • Insight. These findings demonstrated that the custom chatbot was effective in supporting students’ science problem solving, while the general-purpose chatbot led to cognitive offloading.
  • Conclusion. In conclusion, our study found that pedagogy-informed custom GAI chatbots signifi- cantly fostered stronger interactions and more diverse cognitive interactions in science problem solving than the general-purpose chatbot.
EducationResearch articlefull textpreprint / unreviewed
2026-06-30 · Springer Link (Chiba Institute of Technology) · Research article
Lin Miao
  • Insight. Findings reveal that students lacking foundational AI literacy tend to over-rely on AI for analytical tasks, resulting in diminished capabilities in information discernment, logical reasoning, and reflective questioning; conversely, those with basic AI literacy demonstrate more rational utilization of AI tools, enabling deeper collaborative human-machine thinking.
  • Conclusion. Based on these findings, this paper argues that rapid advancements in generative AI lead to cognitive outsourcing behaviors among students, where excessive reliance on AI diminishes independent thinking capacity and fosters information homogenization.
EducationResearch articleabstract onlypublisher record
2026-06-30 · Cognitive Processing · Empirical
Purvi Jain, Deshna Shajit, Ansh Maini, Yasmeen Shahin + 1 more
  • Insight. Although they present potential benefits like increased efficiency and reduced lower-level cognitive load, they are accompanied by serious risks such as over reliance and impacted independent critical reasoning.
  • Conclusion. Overall, we conclude that although AI improves task accuracy and lowers mental effort, it does not necessarily save time or uniformly enhance or reduce critical thinking.
EducationEmpiricalabstract onlypublisher record
2026-07-16 · Higher Education Research & Development · Empirical
Gail Crimmins, Erin Sanders O’Leary, Eileen Strempel
  • Insight. Findings reveal a spectrum of responses, combining enthusiasm for GenAI’s creative affordances with concerns about academic integrity, cognitive dependency, cultural bias, environmental sustainability, and digital inclusion.
  • Conclusion. We argue that resistance to GenAI reflects pedagogical care rather than technological reluctance, and that principled approaches are required to safeguard scholarly rigour, equity, and student agency.
EducationEmpiricalabstract onlypublisher record
2026-05-01 · Computers in Human Behavior Reports · Research article
Kamila Misiejuk, Sonsoles López-Pernas, Rogers Kaliisa, Mohammed Saqr
  • Insight. The findings show how cognitive effort is distributed in the student-AI collaboration and how prompting strategies develop in multiple assignments.
  • Conclusion. Prompting an AI is not a neutral act: it reflects how students think, plan, and offload cognitive effort.
EducationResearch articleabstract onlypublisher record
2026-01-01 · RESEARCH JOURNAL OF PHILOSOPHY & SOCIAL SCIENCES · Research article
Grace Leishiyui Kapai, Keisham Shitaljit Singh
  • Insight. The paper distinguishes AI scaffolding, which reduces extraneous difficulty while retaining learner reasoning, from offloading that also removes the effort needed to build understanding.
  • Conclusion. Its educational boundary is whether students still explain, test, adapt, and eventually perform without the support; accessible metadata does not establish a causal effect.
EducationResearch articleabstract onlypublisher record
2026-08-06 · Advances in computational intelligence and robotics book series · Research article
Ayesha Asghar Gill, Fouzia Janjua, Mehwish Zahoor, Adnan Afzal
  • Insight. Due to the swift growth of artificial intelligence, the cognitive practices have been altered by mass cognitive offloading, casting doubts on the neuropsychological implications of this process in the long term.
  • Conclusion. There are ethical, educative, and social implications that are discussed, as well as the mitigation measures that are advocating balanced human-AI cooperation in order to maintain autonomy and cognitive resilience.
EducationResearch articleabstract onlypublisher record
2026-04-19 · Learning Disability Quarterly · Research article
Yerin Seung, James D. Basham
  • Insight. It synthesizes research on the cognitive challenges experienced by SWLDs, the benefits of GenAI as a compensatory aid, the risks of excessive cognitive offloading, and the cognitive and motivational factors that shape students’ offloading decisions, culminating in an initial conceptual model.
  • Conclusion. The review concludes with implications for future research and practice to support strategic cognitive offloading, highlighting the need to investigate how SWLDs make offloading decisions and how these decisions influence learning in GenAI-supported contexts.
EducationResearch articleabstract onlypublisher record
2026-01-01 · OSF Preprints (OSF Preprints) · Review
Jingwen Wei, Xin Zhang
  • Insight. Cognitive offloading — the use of external resources to reduce internal cognitive processing demands — is not a new phenomenon; however, the emergence of generative AI introduces a qualitatively distinct form of offloading that is interactive, language-based, and capable of performing complex reasoning tasks on behalf of the user.
  • Conclusion. While prior research on cognitive offloading has focused on physical tools (e.g., pen and paper, calculators, smartphones), the implications of offloading to generative AI in educational contexts remain poorly understood.
EducationReviewabstract onlypreprint / unreviewed
2026-02-08 · International Journal of Global Mental Health Innovation Policy Action Culture & Transformation · Research article
Fayize P V
  • Insight. Drawing on recent empirical studies and theoretical perspectives in cognitive psychology and human–computer interaction, this review examines evidence suggesting that frequent delegation of cognitive tasks to AI tools may contribute to cognitive offloading, reduced mental effort, and diminished engagement of executive functions.
  • Conclusion. Ethical and psychological implications of AI dependency are discussed, emphasizing the need for balanced, human-centered AI use that preserves cognitive autonomy.
EducationResearch articleabstract onlypublisher record
2026-05-29 · International journal on research and development. A management review · Research article
Aarchana Patil, Shraddha Shinde
  • Insight. For example, convenience sampling and the small number of participants used may result in biased results since the generalizability of the findings is questionable.
  • Conclusion. This paper provides various theoretical use and implication to literature on the subject matter, in which cognitive offloading and mental laziness are correlated in a real organizational context.
EducationResearch articlefull textpublisher record
2026-04-10 · OSF Preprints (OSF Preprints) · Review
Obaid Azeem
  • Insight. Findings will inform evidence-based guidelines for balancing AI's educational benefits with risks to cognitive development and academic integrity.
  • Conclusion. This meta-analysis synthesizes empirical evidence on the psychological mechanisms linking student-initiated generative AI use to learning outcomes in secondary and higher education.
EducationReviewabstract onlypreprint / unreviewed
2026-01-01 · International Journal of Multidisciplinary Research and Growth Evaluation · Empirical
Papia Sarkar
  • Insight. The results indicate that mental effort and information overload significantly increase cognitive offloading, while decision fatigue negatively influences it.
  • Conclusion. This study examines the impact of cognitive load dimensions—mental effort, information overload, and decision fatigue—on academic performance in AI-enabled learning environments, with cognitive offloading as a mediating mechanism and Generative AI as a moderating factor.
EducationEmpiricalabstract onlypublisher record
2026-01-01 · SSRN Electronic Journal · Conceptual
Doan Winkel
  • Insight. Generative artificial intelligence (GenAI) tools, particularly conversational large language models such as ChatGPT, are rapidly transforming management education, yet we lack theoretical clarity about why the quality of students' engagement with these tools matters more than mere access.
  • Conclusion. I advance four propositions that specify the mechanisms through which trained GenAI engagement enhances opportunity recognition, business model development, and customer discovery.
EducationConceptualabstract onlypreprint / unreviewed
2026-07-02 · Frontiers in Educational Practice and Research · Research article
Meena Jha, Stephen Colbran, K. Purnell
  • Insight. While these technologies offer accessibility, personalisation, and educational support, they raise concerns regarding cognitive offloading, eroding biological memory, and development of human expertise.
  • Conclusion. This paper argues that prompting AI systems should not be conflated with learning because expertise depends upon durable internal cognitive structures developed through encoding, retrieval, schema formation, and critical reflective reasoning over time.
EducationResearch articleabstract onlypublisher record
2026-01-01 · SSRN Electronic Journal · Empirical
Bruno Campello de Souza, Agostinho Serrano de Andrade Neto, Antônio Roazzi
  • Insight. These findings are interpreted through a dual-process framework distinguishing passive cognitive offloading from active cognitive integration.
  • Conclusion. The Sophotechnic Mediation Scale demonstrated high internal consistency (α = .941) and unidimensionality.
EducationEmpiricalabstract onlypreprint / unreviewed
2026-03-31 · Journal of Professional & Applied Psychology · Review
Fareeha Nayab, Fozia Bibi, Asma Mukhtar
  • Insight. Results also show problems in the area of 'cognitive offloading', limited critical thinking, cheating in academic work and behavioral dependency.
  • Conclusion. A review of literature on the Technologies Acceptance Model and the Cognitive Offloading Theory led to the proposed Assistance–Reliance–Dependency Continuum, which highlights the dynamic relationship between the student and AI chatbots and includes practical implications for the responsible use of AI in higher education.
EducationReviewfull textpublisher record
2026-01-01 · OSF Preprints (OSF Preprints) · Research article
Eduardo Gamero-Casado
  • Insight. Second, from the available empirical evidence: the studies of Kosmyna et al. (2025), Gerlich (2025), and Yang et al. (2025) document significant negative effects of generative AI use on neural connectivity, critical thinking capacity, and higher-order reasoning in educational settings.
  • Conclusion. This article examines the implications of this phenomenon from three complementary perspectives.
EducationResearch articleabstract onlypreprint / unreviewed
2026-06-15 · Center for Open Science · Research article
Eduardo Gamero Casado
  • Insight. Second, from the available empirical evidence: the studies of Kosmyna et al. (2025), Gerlich (2025), and Yang et al. (2025) document significant negative effects of generative AI use on neural connectivity, critical thinking capacity, and higher-order reasoning in educational settings.
  • Conclusion. This article examines the implications of this phenomenon from three complementary perspectives.
EducationResearch articleabstract onlypreprint / unreviewed
2026-01-01 · Open MIND · Empirical
Emma Di Modica, Margot J. van der Goot
  • Insight. Data will be collected through semi-structured in-depth interviews with habitual GenAI users enrolled at universities in the Netherlands, complemented by a metaphor elicitation task and a counterfactual writing task designed to surface real-time evidence of authorial agency and cognitive dependency.
  • Conclusion. This qualitative study examines how habitual use of generative AI tools, primarily ChatGPT, shapes university students' academic communication practices, their sense of authorial agency over their own writing and thinking, and their experience of themselves as academic communicators over time.
EducationEmpiricalabstract onlypreprint / unreviewed
2026-02-01 · IOSR Journal of Research & Method in Education (IOSRJRME) · Empirical
Trang Hong Nguyen
  • Insight. However, the results also indicate that frequent AI use is associated with higher cognitive offloading, weaker verification behavior, and lower levels of independent critical thinking.
  • Conclusion. Therefore, the findings should be understood as relating to general patterns of AI use, not to the performance of any one tool.
EducationEmpiricalfull textpublisher record
2026-01-01 · UTS Figshare · Research article
Poulton Lodge, Patricia K Leslie, AM Loble
  • Insight. This cognitive offloading from human to AI is especially risky for school students (‘novice’ learners who are building foundational knowledge and skills) when they turn to AI as a tempting substitute, not an amplifier, increase their dependency on the tool and lose access to deeper learning and critical thinking capabilities.
  • Conclusion. It includes specific recommendations for policy and teaching and learning strategies.
EducationResearch articleabstract onlypublisher record
2026-02-24 · International Conference on Artificial Intelligence in Education · Research article
Rahul Kumar, Samita Sarkar
  • Insight. The rapid evolution of AI capabilities means current findings may prove transitional as technology advances.
  • Conclusion. Future empirical research should examine long-term cognitive effects of AI reliance on administrators, stakeholder trust implications when AI-mediated communications are detected, differential equity impacts across diverse school communities, cross-cultural implementation patterns, and effectiveness of hybrid governance approaches for AI integration in educational leadership.
EducationResearch articleabstract onlypublisher record
2026-07-20 · Advances in Respiratory Medicine · Empirical
G. Marchi, Lorenzo Corbetta
  • Insight. In an international survey of 118 expert interventional pulmonologists, 73% expressed concern about procedural skill erosion and 83% about inhibited trainee upskilling, even before widespread clinical AI deployment.
  • Conclusion. Respondents strongly supported AI-free training, simulation, longitudinal monitoring, and minimum unaided procedure volumes, while the study notes that perceived risk is not yet evidence of measured competence loss.
EducationEmpiricalabstract onlypublisher record
2026-01-01 · Open MIND · Research article
Kantesha Sanningammanavara
  • Insight. This registered project proposes a sequential mediation model linking AI use, cognitive offloading, and academic performance among rural undergraduates.
  • Conclusion. Because the accessible OSF record describes the planned model rather than completed results, it establishes scope and hypotheses but no verified performance effect.
EducationResearch articleabstract onlypreprint / unreviewed
2026-07-16 · Zenodo (CERN European Organization for Nuclear Research) · Conceptual
Sudhakar Geruganti
  • Insight. This balanced approach supports sustainable innovation and reduces the risks associated with overdependence on intelligent technologies.
  • Conclusion. Ultimately, this research argues that the future of Artificial Intelligence should not be measured solely by computational performance or automation capabilities, but by its ability to enhance human potential, strengthen cognitive development, promote ethical decision-making, and contribute to sustainable societal progress.
EducationConceptualabstract onlypreprint / unreviewed
2026-07-01 · AI in Clinical Medicine · Research article
E. Lim, Chi Eung Danforn Lim
  • Insight. The review distinguishes 'AI brain fog'—under-engagement, reduced recall, and habitual offloading—from 'AI brain fry'—fatigue caused by excessive outputs, alerts, monitoring, and verification; the two can occur together.
  • Conclusion. These are educational constructs, not diagnoses. The authors propose evaluating healthcare AI by whether it preserves independent judgment, attention, diagnostic humility, and professional accountability as well as saving time.
EducationResearch articlefull textpublisher record
2026-07-23 · AI Policy Exchange Forum · Research article
Marit Brademann
  • Insight. While cognitive offloading is a natural part of human learning, GenAI’s ease of use risks disrupting efficient learning processes, potentially depriving children and youth of their right to quality education, including foundational skills like literacy, numeracy, and critical thinking.
  • Conclusion. The paper concludes that arbitrary cognitive offloading via GenAI jeopardizes the normative goals of quality education, as defined by frameworks like UNICEF’s and Bloom’s Revised Taxonomy.
EducationResearch articleabstract onlypublisher record
2026-08-15 · Journal of the Association for Information Systems · Empirical
Felipe Bravo, Noah Lepore, Adelisa Pelinkovic, Jillian Rossman + 1 more
  • Insight. GenAI tools are transforming how students approach academic work, raising concerns about reduced cognitive engagement.
  • Conclusion. This study introduces the phenomenon of anticipatory cognitive offloading, proposing that students reallocate effort during preparation when they expect AI assistance during assessment.
EducationEmpiricalabstract onlypublisher record
2026-06-30 · Korean Society for Creativity Education · Research article
Yongmin Lee
  • Insight. Analyzing 97 students across three courses with varying creativity demands, the findings are as follows.
  • Conclusion. Based on these findings, the study proposes a “Reverse Human-AI Collaboration Model,” which delays AI intervention during divergent thinking while employing AI as a critical reviewer during convergent stages, and presents educational implications for establishing a human-AI co-learning ecosystem in higher education.
EducationResearch articleabstract onlypublisher record
2026-04-15 · International Educational Review · Empirical
SA Sony, K. Mus
  • Insight. It then considers implications for instructional design, AI system development, and assessment strategies, and outlines directions for empirical research that build on, rather than duplicate, recent experimental findings.
  • Conclusion. While AI tools enhance accessibility, they also introduce cognitive risks by encouraging learners to bypass critical thinking.
EducationEmpiricalabstract onlypublisher record
2026-02-07 · Artificial Intelligence in Lifelong and Life-Course Education · Empirical
Andini Noviyanti Fitriani, Rezky Risaldy, Annajmi Rauf, Shera Afidatunisa
  • Insight. Academic dependency and cognitive offloading also show positive effects, suggesting that dependency can act as scaffolding and offloading can support higher-order thinking when applied strategically and ethically.
  • Conclusion. Future studies should employ longitudinal or experimental designs, include behavioral indicators (e.g., usage logs), and test moderators such as self-regulation, epistemic beliefs, and task complexity to clarify when dependency and offloading shift from supportive to harmful.
EducationEmpiricalfull textpublisher record
2026-05-01 · Computers in Human Behavior Artificial Humans · Research article
Yali Jiang, Haiping Chen
  • Insight. Results indicate that the perceived value of AI is positively associated with both types of cognitive offloading through AI-usage capability.
  • Conclusion. Artificial intelligence (AI) tools, such as ChatGPT, are becoming increasingly embedded in higher education for assisting students in tasks ranging from information retrieval to essay writing and problem solving in China.
EducationResearch articleabstract onlypublisher record
2026-01-20 · Frontiers in Artificial Intelligence · Empirical
Mariza Tsakalerou, Saltanat Akhmadi, Aruzhan Balgynbayeva, Yerdaulet Kumisbek
  • Insight. Participants' self-perceptions of creativity remained stable across both conditions, highlighting the risk of cognitive offloading, where reliance on AI may reduce genuine creative engagement while masking deficits through inflated confidence.
  • Conclusion. The study argues for process-aware and ethically grounded frameworks that balance augmentation with human agency, supporting exploration without eroding the foundations of creative problem-solving.
EducationEmpiricalfull textpublisher record
2026-04-25 · SEET@ICSE · Empirical
Eduardo Oliveira, Michael Fu, Patanamon Thongtanunam, Sonsoles López-Pernas + 1 more
  • Insight. We contribute: (i) an in-workflow design for an AI reviewer that scaffolds learning while mitigating cognitive offloading; (ii) a repeated cross sectional comparison across two cohorts in authentic settings; (iii) a mixed-methods analysis combining objective GitHub metrics with student self-reports; and (iv) evidence-based pedagogical recommendations for responsible, student-led AI-assisted review.
  • Conclusion. Code review is central to software engineering education but hard to scale in capstone projects due to tight deadlines, uneven peer feedback, and limited prior experience.
EducationEmpiricalfull textpublisher record
2026-02-04 · Open MIND · Research article
Lucile Favero, Juan Antonio Pérez-Ortiz, Tanja Käser, Nuria Oliver
  • Insight. The integrative framework links four dimensions—cognition, agency, emotional well-being, and ethics—showing how offloading, reduced autonomy, disengagement, and surveillance practices can reinforce one another.
  • Conclusion. Educational AI should be judged by whether design and governance preserve effortful reasoning, learner agency, meaningful social interaction, and civic capacities, not only immediate learning efficiency.
EducationResearch articlefull textpreprint / unreviewed
2026-04-30 · Journal of Educational Studies · Research article
Atti Yudiernawati, Soetam Rizky Wicaksono, Tavip Dwi Wahyuni
  • Insight. The results show that, at the initial stage, most students perceived AI as a tool that simplifies tasks, leading to reduced cognitive effort and increased dependency.
  • Conclusion. This research concludes that AI can both weaken and strengthen students’ self-efficacy, depending on how it is integrated into the learning process.
EducationResearch articlefull textpublisher record
2026-06-15 · Artificial Intelligence in Lifelong and Life-Course Education · Empirical
Boonchat Mekkaeo, Parkpoom Sungchuay, Supawadee Thibai
  • Insight. Although generative AI improved students’ immediate task performance, the findings show that AI-assisted achievement did not always correspond to independently demonstrable linguistic proficiency.
  • Conclusion. Future research should test these findings using stronger pre-test/post-test and longitudinal designs across more diverse educational contexts.
EducationEmpiricalfull textpublisher record
2026-03-08 · arXiv (Cornell University) · Conceptual
Eduardo Davalos, Yike Zhang
  • Insight. We introduce the Learning Visibility Framework, grounded in three principles: clear specification and modeling of acceptable AI use, recognition of learning processes as assessable evidence alongside outcomes, and the establish- ment of transparent timelines of student actions.
  • Conclusion. Drawing on research in cognitive offloading and learning analytics, we argue that responsible AI integration requires advocating transparency to student ac- tivity.
EducationConceptualfull textpreprint / unreviewed
2026-03-13 · ACADEMIA International Journal for Social Sciences · Empirical
Naheed Akhtar, Uzma Mazhar, Nimra Shehzadi, Hajra Dildar + 2 more
  • Insight. Findings suggest that unregulated AI reliance compromises foundational academic competencies, but structured pedagogical scaffolding can mitigate adverse outcomes, informing evidence-based AI integration policies in higher education.
  • Conclusion. This mixed-methods longitudinal study investigated how generative AI dependency influences memory retention, critical thinking, and decision-making autonomy among 412 university students.
EducationEmpiricalabstract onlypublisher record
2026-10-06 · IGI Global Scientific Publishing · Research article
Doni Sugianto Sihotang, Usep Suhud, Agung Wahyu Handaru
  • Insight. This chapter analyzes the psychological and cognitive consequences of heightened dependence on intelligent technologies, emphasizing dependency and cognitive job delegation in human-technology interactions.
  • Conclusion. This chapter aims to examine how intelligent systems transform cognitive processes, affect critical thinking, and contest conventional human-centered design principles.
EducationResearch articleabstract onlypublisher record
2026-05-28 · arXiv (Cornell University) · Conceptual
Clayton Cohn, Surya Rayala, Siyuan Guo, Hanchen David Wang + 12 more
  • Insight. These findings position theory-guided, multimodal LLM agents as a promising path toward classroom AI integration that amplifies students' reasoning rather than replacing it.
  • Conclusion. LLM pedagogical agents are proliferating, yet recent findings have raised questions about their adherence to established theories of learning and, by extension, their educational value.
EducationConceptualfull textpreprint / unreviewed
2026-08-15 · Journal of the Association for Information Systems · Research article
Lauren De Guzman, Timothy R. Hill, Yu Chen
  • Insight. The findings suggest that collaborative AI-supported problem-solving can remain useful to students while also encouraging greater engagement and critical thinking throughout the learning process.
  • Conclusion. As generative AI becomes more prevalent in higher education, simply limiting student use might not be sufficient to address the risks of cognitive offloading.
EducationResearch articleabstract onlypublisher record
2026-07-20 · International Journal of Technology in Education · Empirical
Jamilah Mohammed Alamri, Waleed Salim Alhalafawy
  • Insight. The findings suggest that GenAI is more effective at supporting retention when used to scaffold the active reasoning of learners as opposed to substituting for their cognitive processing.
  • Conclusion. Even though the effects of Generative AI (GenAI) on learning retention are still not adequately understood, its use in higher education is increasing.
EducationEmpiricalabstract onlypublisher record
2025-05-14 · 2025 5th International Conference on Artificial Intelligence and Education (ICAIE) · Research article
Garth H. Elzerman
  • Insight. The findings highlight the need for a balanced approach in AI-assisted education, ensuring that AI serves as a tool to enhance rather than replace learners' cognitive development.
  • Conclusion. By drawing on cognitive theories, empirical studies, and case study evidence, this paper explores these cognitive drawbacks and their implications for AI design and educational practice.
EducationResearch articleabstract onlypublisher record
2025-08-15 · arXiv · Research article
Rénald Gesnot
  • Insight. This first chapter has explored the con- ceptual foundations of AI and human cognition and highlighted the main mechanisms at play—offloading of cognitive load, transactive memory, changes in attentional engage- ment, etc.
  • Conclusion. The analysis reveals a dialectical tension between augmentation and decline, which must be managed through thoughtful and measured use of AI technologies.
EducationResearch articlefull textpreprint / unreviewed
2025-01-01 · ASEAN Journal of Psychiatry · Conceptual
Christopher Cleverly
  • Insight. Drawing on evidence such as the impact of AI on 40% of global jobs and cognitive offloading in education, we challenge the notion that technical fixes alone can mitigate these effects.
  • Conclusion. This research introduces a framework to elucidate how automation bias in Large Language Models (LLMs) amplifies biases through human over-reliance, leading to critical thinking atrophy and the propagation of biases into human cognition and social systems.
EducationConceptualfull textpublisher record
2025-07-15 · Annals of King Edward Medical University · Empirical
Mashaal Sabqat, Noorul Ain, Sana Iqbal, Rehan Ahmed Khan
  • Insight. The initial 47-item questionnaire was refined to 36 items, with an S-CVI/Ave of 0.88 and a CCA of 90%.
  • Conclusion. The increasing reliance on AI among medical students is associated with metacognitive laziness, emphasizing the need for careful AI integration to promote independent learning and critical thinking.
EducationEmpiricalfull textpublisher record
2025-10-03 · Center for Open Science · Research article
Nikhil Jain, M. C. Kiran
  • Insight. This conceptual paper argues for reformthrough three interconnected lenses—pedagogical transformation, curriculum redesign, andinstitutional adaptation, grounded in the theories of cognitive offloading and critical pedagogy.Drawing on both historical parallels and contemporary research, we demonstrate that generativeAI, when thoughtfully integrated, can support rather than undermine learning, creativity, andstudent agency.
  • Conclusion. We conclude by outlining future research directions to betterunderstand how generative AI is reshaping cognitive development, student learning, and theinstitutional structures of education.
EducationResearch articleabstract onlypreprint / unreviewed
2025-06-10 · Proceedings. · Empirical
Bookyung Shin, Min Ju Kang, Seunmin Eun, Young Hoan Cho
  • Insight. Process traces from 47 South Korean higher-education students showed that translation annotations supported more linear, purposeful reading, while deep-reasoning questions prompted more complex integration of information.
  • Conclusion. Different annotations distributed cognition differently: translation primarily offloaded processing, whereas questions more often extended and reorganized the reader's own reasoning.
EducationEmpiricalabstract onlypublisher record
2025-12-05 · Frontiers in Psychology · Research article
Binny Jose, Deepak Joseph, Visakh V. Mohan, Elizabeth Alexander + 2 more
  • Insight. In addition, future empirical studies will be able to operationalize the levels of off-loading, to help define at what point in time each of the three levels of off-loading result in an increase in task completion, but also an increase in cognitive costs, and to ultimately guide educators in designing responsible and effective technologies for supporting learners.A visual overview of the proposed taxonomy is presented in Figure 1 The figure illustrates the continuum of cognitive offloading from assistive to disruptive forms, showing how increasing external automation progressively reduces internal control.
  • Conclusion. Therefore, future research needs to assess how these seemingly minor design decisions affect cognitive processes in educational and health care environments, where internal agency is most important.In order to diminish reliance on AI, all future psychological interventions will focus on utilizing internal cognitive approaches that support a variety of cognitive processes.
EducationResearch articlefull textpublisher record
2025-06-30 · Studies in English Education · Empirical
Yun Hee Chung, Byung Kyoo Ahn
  • Insight. These results call for explicit, sustained strategy instruction to mitigate cognitive offloading and to cultivate autonomous, critically engaged writers within AI-mediated academic environments.
  • Conclusion. Cluster analysis identified three distinct learner profiles (Cluster 1, Cluster 2, and Cluster 3), pointing to the need for differentiated scaffolding.
EducationEmpiricalabstract onlypublisher record
2025-12-02 · Journal of English Studies in Arabia Felix · Research article
Marwan Saeed Saif Moqbel
  • Insight. The findings are synthesized into a theoretically grounded framework that explains AI’s impact on the breadth and depth of language education, particularly in contexts where English is not used as a native language.
  • Conclusion. Drawing on a comprehensive review of empirical studies through pedagogical and SLA theoretical lenses, the study looked into the pedagogical implications of the complex relationship between AI-driven personalization and curricular narrowing in language education.
EducationResearch articleabstract onlypublisher record
2025-12-31 · Cultura Ciencia y Deporte · Research article
Alejandro Espeso García
  • Insight. A fundamental dilemma thus arises: is GenAI an amplifier that democratizes performance or a mechanism of dependence that leads to cognitive offloading?
  • Conclusion. Drawing on recent literature, this article examines its educational, cultural, and social implications, highlighting the need to understand it both as a technical tool and as a phenomenon that transforms the relationship between mind and knowledge.
EducationResearch articleabstract onlypublisher record
2025-05-13 · Scientific Reports · Empirical
Javed Iqbal, Zarqa Farooq Hashmi, Muhammad Zaheer Asghar, Muhammad Naseem Abid
  • Insight. This may affect the generalizability of the findings to students in other contexts.
  • Conclusion. Limitations and future research Our study also has some limitations that should be considered when the results are interpreted.
EducationEmpiricalfull textpublisher record
2025-11-24 · OSF Preprints (OSF Preprints) · Empirical
Florina Pînzaru, Alexandra Zbuchea, Sebastian Fitzek
  • Insight. AI-ECOA / SNSPA — Mixed-Methods Protocol & Survey Package From Cognitive Outsourcing to AI-Ready Skills: Evidence from a Mixed-Methods Study in a Digital Labour Market Protocol version: 29 Sep 2025 | Lead: Sebastian Fitzek (SNSPA) | Project: AI-ECOA Executive summary This protocol specifies a short, high-impact mixed-methods design: a cross-sectional online survey (~8–10 minutes) with two performance-based micro-tasks (prompt literacy quiz; score 0–2), triangulated with prior AI-ECOA focus-group insights.
  • Conclusion. Analysis plan (abridged) Descriptives; reliability (α/ω); EFA/CFA as needed; hierarchical regressions with FDR control (q=.05); mediation (bootstrap, 5,000).
EducationEmpiricalabstract onlypublisher record
2025-12-13 · arXiv (Cornell University) · Empirical
Amir Yunus, Peng Rend Gay, Oon Teng Lee
  • Insight. These findings suggest that Generative AI's educational influence extends beyond instructional efficiency to shape cognitive engagement, self-regulation, and learner equity.
  • Conclusion. We conclude that while Generative AI can substantially enhance the teaching experience, achieving meaningful learning gains requires rigorous attention to learner behaviour and the equitable design of AI-supported environments.
EducationEmpiricalfull textpreprint / unreviewed
2025-11-30 · FOSTER Journal of English Language Teaching · Empirical
Muhammad Mobeen Tahir, Andi Sahtiani Jahrir
  • Insight. However, a significant negative correlation between cognitive offloading and reading habits (r = –0.54) suggests that increased dependence on ChatGPT reduces students’ motivation for traditional reading.
  • Conclusion. The results underscore the need for balanced AI integration that promotes critical reading, reflective thinking, and responsible technology use in postgraduate education.
EducationEmpiricalabstract onlypublisher record
2025-08-22 · OpenAlex · Empirical
Hui Hong
  • Insight. A mixed-method study comparing ICO students and controls showed that ICO significantly improved critical thinking skills, especially in the depth of argument and evidence synthesis.
  • Conclusion. Implications suggest that the ICO framework effectively reduces the risk of cognitive surrogacy through structured collaboration and provides a viable model for using artificial intelligence to augment basic intellectual skills.
EducationEmpiricalabstract onlypublisher record
2025-02-05 · Journal of Ecohumanism · Research article
Hui Hong, Chantana Viriyavejakul, Poonsri Vate-U-Lan
  • Insight. The findings contribute to the growing body of literature on the role of AI in educational settings, highlighting the potential of AI-assisted cognitive offloading to advance critical thinking and writing skills through structured instructional design.
  • Conclusion. This study explores the integration of Generative AI technology for cognitive offloading in the context of English essay writing, with the aim of promoting inclusive and equitable quality education.
EducationResearch articleabstract onlypublisher record
2025-12-18 · arXiv (Cornell University) · Research article
Giovanni Adorni
  • Insight. The paper proposes Cyber Humanism in Education around reflexive competence, algorithmic citizenship, and dialogic design, treating educators and learners as co-authors of AI-rich learning infrastructures.
  • Conclusion. Prompt-based learning and educator certification can strengthen epistemic agency, but workload, equity, and governance tensions require AI to be managed as part of the learning environment rather than a neutral external tool.
EducationResearch articlefull textpreprint / unreviewed
2025-12-20 · Colloquia Academic Journal of Culture and Thought · Research article
Vincent J. Hooper
  • Insight. These findings carry significant normative implications across three domains.
  • Conclusion. For reasoning practices, professionals in law, medicine, and finance should adopt structured protocols that mandate independent analysis prior to consulting AI recommendations.
EducationResearch articlefull textpublisher record
2025-12-08 · International Multidisciplinary Research Journal · Empirical
OMI, Ph.D, DLitt (hc), PDTQM FRIEdr, FRIM 4, Marlita V. Madera, Andrea Ong, Cedric Jon R. Solivio + 2 more
  • Insight. Correlation results showed no significant relationship between AI dependency and retention, and no significant relationship between AI dependency and deep learning.
  • Conclusion. Future research should assess AI’s long-term impacts on creativity, critical thinking, and knowledge retention to safeguard learning.
EducationEmpiricalabstract onlypublisher record
2025-09-15 · IEEE · Research article
Vladimir N. Azarov, Vladimir V. Gorokhov
  • Insight. This article analyzes the critical risks associated with the widespread adoption of artificial intelligence (AI) technologies in daily life and professional activities, particularly the phenomenon of “cognitive offloading” and its potential transition into the atrophy of fundamental cognitive skills.
  • Conclusion. The mechanisms for the formation of new forms of technological dependence on AI systems are examined.
EducationResearch articleabstract onlypublisher record
2025-11-21 · Open MIND · Empirical
Thum, Elias B., Papenmeier, Frank
  • Insight. Together, the findings extend research on cognitive offloading to the context of generative AI, suggesting that even minor interface parameters, such as typing speed, can alter the balance between internal and external strategies.
  • Conclusion. Generative artificial intelligence (AI) is increasingly integrated into everyday problem solving, yet still too little is known about how system characteristics and user abilities shape cognitive offloading to such tools.
EducationEmpiricalabstract onlypreprint / unreviewed
2025-07-15 · Forum for Linguistic Studies · Empirical
Hui Hong, Poonsri Vate-U-Lan, Chantana Viriyavejakul
  • Insight. Results revealed that the AI-enabled cognitive offload group demonstrated significantly greater improvements in standardized critical thinking assessments and produced higher-quality essays in terms of logical coherence, evidence use, and originality.
  • Conclusion. It offers practical implications for educators seeking to incorporate AI tools without compromising the development of essential cognitive skills, proposing that structured cognitive offload instruction can serve as an effective approach to fostering critical thinking in second-language writing contexts.
EducationEmpiricalfull textpublisher record
2025-12-02 · Clareus Scientific Publications · Empirical
Vivian, Robin
  • Insight. Coding with ChatGPT: Empirical Evidence of Cognitive Offloading in Computer Science Education https://clareus.org/csse 19 9.
  • Conclusion. The risk is not ignorance, but intellectual complacency — a world where knowing is replaced by generating.
EducationEmpiricalfull textpublisher record
2025-06-30 · arXiv (Cornell University) · Empirical
Georgios P. Georgiou
  • Insight. Together, the results suggest that AI tools may encourage students to become “lazy thinkers”, potentially reducing their cognitive engagement when relying heavily on tools like ChatGPT.
  • Conclusion. Larger and more diverse participant groups, as well as varied academic tasks and AI tools, should be explored in subsequent research to validate and extend these results.
EducationEmpiricalfull textpreprint / unreviewed
2025-12-15 · AI Brain and Child · Conceptual
Sisi Tao
  • Insight. The perspective proposes five age bands for K–12 GenAI use, ranging from no direct use at ages 3–5 to progressively more independent, collaborative, and analytical use as executive and evaluative capacities mature.
  • Conclusion. Its age cutoffs are precautionary, falsifiable design guidelines rather than validated clinical thresholds; the author calls for empirical testing and adaptation across contexts while prioritizing student agency.
EducationConceptualfull textpublisher record
2025-07-10 · VOCATIONAL: Jurnal Inovasi Pendidikan Kejuruan · Empirical
Hafid Rahmandan, M. R. Badu, Andi Maga Umara
  • Insight. The findings reveal that students utilize AI as a multifunctional tool to enhance academic efficiency.
  • Conclusion. It is recommended that higher education institutions design curricula that foster critical and ethical understanding of technology, as well as encourage academic tasks based on analysis and originality.
EducationEmpiricalabstract onlypublisher record
2025 · Elsevier BV · Research article
Yali Jiang, Chen Haiping
  • Insight. Across scale-development samples and a study of 667 Chinese university students, perceived AI value predicted both lower- and higher-order offloading through AI-use capability; strong evaluation capability weakened the lower-order pathway but not the higher-order one.
  • Conclusion. AI literacy is not one skill: operating competence can increase delegation, while evaluation competence selectively constrains it, leaving higher-order cognitive delegation as the harder risk to regulate.
EducationResearch articleabstract onlypreprint / unreviewed
2025-01-01 · Open Science Framework · Empirical
Fitzek, Sebastian, Pînzaru, Florina, Zbuchea, Alexandra
  • Insight. This preregistration documents the AI-ECOA study examining links between students’ AI use (AI_USE), prompt self-efficacy (PSE), AI literacy (AI_LIT), ethics/regulation awareness (ETHICS), social-listening (SL) and chatbot design readiness (CHATBOT), transversal soft skills (SOFT), and AI-ready career preparedness (READINESS).
  • Conclusion. Analysis plan: descriptives; reliability (α/ω); EFA/CFA as needed; hierarchical regressions with FDR control (q=.05); mediation with bootstrap (5,000 resamples); sensitivity checks excluding speeders and cases with >50% missing in core scales.
EducationEmpiricalabstract onlypublisher record
2025-11-10 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Needham, Eric
  • Insight. This paper presents a developmental risk framework for understanding how AI-mediated environments alter cognitive, emotional, and identity formation processes in children and adolescents.
  • Conclusion. This is a theoretical synthesis intended to clarify developmental mechanisms and guide future research, educational practice, and system design.
EducationResearch articleabstract onlypreprint / unreviewed
2025-01-03 · Societies · Empirical
Michael Gerlich
  • Insight. The findings revealed a significant negative correlation between frequent AI tool usage and critical thinking abilities, mediated by increased cognitive offloading.
  • Conclusion. This study contributes to the growing discourse on AI’s cognitive implications, offering practical recommendations for mitigating its adverse effects on critical thinking.
EducationEmpiricalabstract onlypublisher record
2025-12-31 · Open MIND · Review
Ashwini Jadhav Ashwini Jadhav
  • Insight. The findings showed that AI tools have both advantages and disadvantages from offloading.
  • Conclusion. The use of artificial intelligence (AI) tools is rapidly increasing among college students.
EducationReviewabstract onlypublisher record
2024-12-01 · IMCC Journal of Science · Research article
Maria Rizalie Lindo, Cherry Ann Cutad
  • Insight. Structured generative AI use was associated with higher posttest self-regulated learning and substantially higher cognitive offloading.
  • Conclusion. Generative artificial intelligence (AI) is rapidly reshaping how senior high school students study, yet its implications for self-regulated learning and learning durability remain poorly resolved in authentic classrooms.
EducationResearch articleabstract onlypublisher record
2024-12-10 · British Journal of Educational Technology · Empirical
Yizhou Fan, Luzhen Tang, Huixiao Le, Kejie Shen + 5 more
  • Insight. The results revealed that: (1) learners who received different learning support showed no difference in post‐task intrinsic motivation; (2) there were significant differences in the frequency and sequences of the self‐regulated learning processes among groups; (3) ChatGPT group outperformed in the essay score improvement but their knowledge gain and transfer were not significantly different.
  • Conclusion. Implications for practice and/or policy When using AI in learning, learners should focus on deepening their understanding of knowledge and actively engage in metacognitive processes such as evaluation, monitoring, and orientation, rather than blindly following ChatGPT's feedback solely to complete tasks efficiently.
EducationEmpiricalabstract onlypublisher record
⚙  Work & expertise (70)
2026-08-15 · Journal of the Association for Information Systems · Empirical
Larry Zhiming Xu, Gabriel Velez, Jacklynn Fitzgerald, Jungmin Lee + 1 more
  • Insight. Preliminary results show that continuous AI access significantly improved quantitative performance, while verbal task performance did not differ across conditions.
  • Conclusion. Drawing on theories of cognitive offloading, technology dependence, and expectation violation, we argue that AI disruption does not simply return users to an unaided baseline.
Work & expertiseEmpiricalabstract onlypublisher record
2026-06-14 · Journal of the Association for Information Systems · Empirical
Jacopo Valentini, Aurelio Ravarini
  • Insight. This study investigates how trust, shaped by perceived social presence, contributes to higher-order cognitive offloading—the tendency to shift critical thinking to Generative AI chatbot tools.
  • Conclusion. The adoption of Generative AI chatbots in organizations is transforming how employees perform cognitive work, enhancing efficiency and supporting augmentation.
Work & expertiseEmpiricalabstract onlypublisher record
2026 · Elsevier BV · Research article
Bruce Rishel
  • Insight. This reduces the information processing cost of integrity, allowing agents to "satisfice" safely without the "moral wiggle room" (Dana et al., 2007) that currently facilitates systemic defection.
  • Conclusion. We argue that the "Trust Tax" is a product of cognitive overload, and that shifting from ceremonial verification to algorithmic shared destiny aligns economic incentives with the limits of human cognition.
Work & expertiseResearch articleabstract onlypreprint / unreviewed
2026-01-21 · arXiv · Research article
Oleg Romanchuk, Roman Bondar
  • Insight. We show that this is not a process deviation or technical defect, but a structural property of deployments where decision generation throughput exceeds bounded human verification capacity.
  • Conclusion. We conclude that unless organizations explicitly redesign decision boundaries or reassign responsibility away from individual decisions toward batch- or system-level ownership, responsibility vacuum remains an invisible but persistent failure mode in scaled agent deployments.
Work & expertiseResearch articlefull textpreprint / unreviewed
2026-02-01 · arXiv.org · Empirical
Mohan Reddy
  • Insight. Building on Woods (1984) Keyhole Effect, the cognitive cost of viewing large information spaces through narrow viewports, I show that chat interfaces systematically degrade analytical performance through five mechanisms: (1) constant content displacement defeats hippocampal spatial memory systems; (2) hidden state variables exceed working memory capacity (approximately 4 chunks under load); (3) forced verbalization triggers verbal overshadowing, degrading visual pattern recognition; (4) linear text streams block epistemic action and cognitive offloading; (5) serialization penalties scale with data dimensionality.
  • Conclusion. The paper concludes with falsifiable hypotheses and experimental paradigms for empirical validation.
Work & expertiseEmpiricalfull textpreprint / unreviewed
2026-06-30 · Journal of Computers Mechanical and Management · Empirical
Hongmei Pan, Rasheedul Haque, Sayeeduzzafar Qazi, Ayma Fatima + 1 more
  • Insight. Results show that decision fatigue is driven primarily by persuasion knowledge activation and reduced cognitive offloading, whereas trust calibration does not significantly predict depletion.
  • Conclusion. Short-video platforms such as Douyin and TikTok embed algorithmic product recommendations within entertainment streams, positioning personalization as a facilitator of decision-making.
Work & expertiseEmpiricalfull textpublisher record
2026-04-03 · arXiv (Cornell University) · Empirical
Michael Caosun, Sinan Aral
  • Insight. Our analysis suggests that even with full awareness of skill atrophy, AI productivity may come with potentially hidden impacts such as reduced productivity, the deskilling of augmented workers, and divergence in the worker population.
  • Conclusion. These results do not imply that skill atrophy is always harmful or that AI adoption should be restrained in general.
Work & expertiseEmpiricalfull textpreprint / unreviewed
2026-06-15 · Journal of the Association for Information Systems · Empirical
Anna Sophia Sommer, Julia Seitz Saenger, Alexander Maedche
  • Insight. A laboratory experiment with 101 participants found that a higher-agency anticipatory writing assistant increased cognitive load and reduced critical-thinking performance compared with a lower-agency reflexive assistant.
  • Conclusion. The performance deficit persisted on a later unaided task, suggesting that more agent initiative can create negative spillover rather than simply reduce user effort.
Work & expertiseEmpiricalabstract onlypublisher record
2026-02-22 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Zaelani
  • Insight. The analysis suggests how distinct computational environments and workflows may cultivate specific professional archetypes (The Augmented Creator, The Knowledge Director, and The Researcher), each with characteristic patterns of cognitive trade-off.
  • Conclusion. The paper concludes by examining the structural tensions of this paradigm.
Work & expertiseResearch articleabstract onlypreprint / unreviewed
2026-01-11 · arXiv (Cornell University) · Research article
Andrew D. Maynard
  • Insight. Four mechanisms of potential bypass are identified: processing fluency decoupled from understanding, trust-competence presentation without corresponding stakes, cognitive offloading that delegates evaluation itself to the AI, and optimization dynamics that systematically produce sycophancy.
  • Conclusion. This paper proposes that a significant epistemic risk from conversational AI may lie not in inaccuracy or intentional deception, but in something more fundamental: these systems may be configured, through optimization processes that make them useful, to present characteristics that bypass the cognitive mechanisms humans evolved to evaluate incoming information.
Work & expertiseResearch articlefull textpreprint / unreviewed
2026-01-01 · Open MIND · Research article
Justin N. Wahlers
  • Insight. This preregistered study separates invested effort from demanded effort to test which subjective experience predicts intentions and choices to delegate cognitive work to AI.
  • Conclusion. It is a study registration, so it specifies a testable distinction about effort-driven offloading but does not yet support a results claim.
Work & expertiseResearch articleabstract onlypreprint / unreviewed
2026-05-23 · arXiv (Cornell University) · Research article
Sora Kang
  • Insight. Our field study suggests that this constraint significantly reduced cognitive burden while fostering a residual effect of internalized questioning habits.
  • Conclusion. In creative domains where the labor of articulation is central to the craft, how should we design Tools for Thought that enhance rather than bypass human cognition?
Work & expertiseResearch articlefull textpreprint / unreviewed
2026-02-01 · arXiv (Cornell University) · Research article
Yueqing Hu, Yixuan Jiang, Zehua Jiang, Xiao Wen + 1 more
  • Insight. These findings highlight a critical gap between behavioral modification and authentic value alignment in artificial societies.
  • Conclusion. Using a full factorial design across three state-of-the-art LLMs, we analyzed both behavioral outcomes and internal reasoning chains.
Work & expertiseResearch articlefull textpreprint / unreviewed
2026-06-27 · Brain Sciences · Research article
Wissem Dhahbi, David B. Pyne, Ismail Dergaa, Daniel Zeitouny + 4 more
  • Insight. Our narrative review aimed to (i) synthesize neurobiological evidence on the mechanisms through which task-irrelevant digital interruption impairs deep thinking; (ii) discuss the conditions required for deep thinking and the potential threats posed by contemporary developments, including generative artificial intelligence-related cognitive offloading; and (iii) elaborate evidence-based, multi-level recommendations for research institutions.
  • Conclusion. Direct evidence linking these intervals to specific research-impact outcomes remains limited, and institutional interventions should be prospectively evaluated.
Work & expertiseResearch articleabstract onlypublisher record
2026-05-14 · Center for Open Science · Research article
Tobias Feldmann-Wüstefeld, Marian-David Wolff, Eva Wiese
  • Insight. However, this effect disappeared under high memory load, suggesting that increased cognitive demands limit the extent of offloading.
  • Conclusion. Efficient collaboration between humans and artificial intelligence (AI) systems requires a dynamic allocation of cognitive resources.
Work & expertiseResearch articleabstract onlypreprint / unreviewed
2026-07-21 · Center for Open Science · Research article
Lauren Danny
  • Insight. Findings indicate that participants experience reduced critical synthesis and increased reliance on automated heuristics.
  • Conclusion. These results suggest that organizational policies must prioritize cognitive resilience or human-in-the-loop verification to mitigate the risks of structural deskilling in the age of automated cognition.
Work & expertiseResearch articleabstract onlypreprint / unreviewed
2026-02-22 · arXiv (Cornell University) · Empirical
Sreecharan Sankaranarayanan
  • Insight. Results reveal a collapse of competence: both AI groups significantly outperformed the manual control on functional utility (p < .001) and did not differ from each other (p = .64), yet unrestricted AI users suffered a 77% failure rate on a subsequent 30-minute AI-blackout maintenance task, vs.
  • Conclusion. Without pedagogical guardrails, we argue this is fundamentally misaligned with cognitive skill acquisition.
Work & expertiseEmpiricalfull textpreprint / unreviewed
2026-07-20 · Behavioral Sciences · Research article
Latifah Hamdan Alghamdi, Talal Musaed Alghizzi
  • Insight. Overall, the evidence suggests that the effectiveness of AI-mediated feedback depends not only on the quality of the feedback provided but also on how learners regulate and engage with it during revision.
  • Conclusion. Greater methodological transparency and more rigorous process-tracing designs are needed to strengthen future research on AI-supported writing development.
Work & expertiseResearch articleabstract onlypublisher record
2026-08-07 · Journal of Intelligence · Empirical
Jing Su, Zhuo Wang, Zhen Qiang, Zhonghou Wang
  • Insight. In a structural equation model with bias-corrected bootstrapping, dependency strongly predicted TSC and partially mediated the effect of AI anxiety; impulsivity raised dependency but showed a suppression pattern (a positive indirect effect offset by a null total effect), and the model explained 35% of the variance in the concern.
  • Conclusion. Cognitive offloading—delegating cognitive work to external tools—is basic to human cognition, but generative AI (GenAI) amplifies it radically: entire cognitive products can now be produced on request.
Work & expertiseEmpiricalabstract onlypublisher record
2026-06-04 · arXiv (Cornell University) · Research article
Anna Mikeda
  • Insight. We present a taxonomy of six metacognitive capacities organized by temporal phase, characterize their tendencies under routine AI use, and show how individually rational adaptation produces emergent social costs.
  • Conclusion. Recent studies reveal a paradox: AI enhances individual creative outputs while reducing collective diversity.
Work & expertiseResearch articlefull textpublisher record
2026-05-07 · arXiv (Cornell University) · Conceptual
Xuening Wu, Yanlan Kang, Qianya Xu, Kexuan Xie + 4 more
  • Insight. These results suggest that the trajectory of AI systems is shaped not only by model design, but by the dynamics of human-AI co-evolution.
  • Conclusion. Large language models (LLMs) are reshaping how knowledge is produced, with increasing reliance on AI systems for generation, summarization, and reasoning.
Work & expertiseConceptualfull textpreprint / unreviewed
2026-06-22 · ACM · Research article
Caterina Fregosi, Simon Fischer, Chiara Natali, Hanna Schraffenberger + 1 more
  • Insight. This workshop responds to the observation that, in work contexts, AI systems may promote unintended socio-cognitive consequences, including automation bias, overreliance, reduced critical engagement and sense of agency, and gradual deskilling.
  • Conclusion. Through structured collaborative activities, participants will identify benefits, risks, protective factors, design strategies, and evaluation measures.
Work & expertiseResearch articleabstract onlypublisher record
2026-01-01 · Open MIND · Review
Jorge Pereira Campos
  • Insight. This scoping review systematically maps the empirical evidence on generative AI use and cognitive development in adolescents - the developmental window during which executive function maturation is most rapid and most experience-dependent.
  • Conclusion. The contribution is threefold: (1) documenting the demographic gap in existing AI-cognition research, including the absence of adolescent participants from prominent neuroscience studies (e.g., Kosmyna et al., 2025); (2) identifying methodological patterns that limit current inference in the adolescent-focused literature; and (3) operationalising the cognitive offloading paradox (Risko & Gilbert, 2016, and adjacent literature) as an organising frame for the developmental research questions the next phase must address.
Work & expertiseReviewabstract onlypreprint / unreviewed
2026-07-31 · Journal of Intelligent Decision Making and Information Science · Conceptual
R. Guerrero-Chirinos, M. Medina-Romero, Jean‐Marc Pierre, Fernando Montemayor-Ibarra + 2 more
  • Insight. This conceptual article, oriented toward theory building through theoretical synthesis, integrates three consolidated bodies of literature, cognitive offloading, desirable difficulties, and metacognition, to theorize the tension between delegating cognition to AI and developing self-regulation.
  • Conclusion. The model derives a typology of four uses and seven testable propositions with their evidential status, and identifies implications for pedagogical design, tool design, and research.
Work & expertiseConceptualfull textpublisher record
2026-06-04 · arXiv (Cornell University) · Research article
Alexander Apartsin, Yehudit Aperstein
  • Insight. Schools still measure unaided performance, yet the real task is to produce good work with AI: framing an ill-defined task, judging the output, and steering the model toward a better result.
  • Conclusion. Generative AI makes answers easy and understanding hard, and uncritical use invites cognitive offloading.
Work & expertiseResearch articlefull textpreprint / unreviewed
2026-07-02 · arXiv (Cornell University) · Conceptual
Santosh Premi Adhikari
  • Insight. The proposed model suggests that AI systems can create a loop in which uncertainty reduction, fluent answers, reassurance, and cognitive offloading encourage continued querying, while synthesis and output are delayed.
  • Conclusion. Doom researching should not be understood as a rejection of generative AI.
Work & expertiseConceptualfull textpreprint / unreviewed
2026-04-30 · Korean Institute of Smart Media · Empirical
Kwang Sep Jean, Han-Sol Lee, Ju Young Kang
  • Insight. The key findings are as follows: First, the use of generative AI tools by novice programmers significantly increased cognitive offloading.
  • Conclusion. The purpose of this study is to empirically investigate the impact of the proliferation of generative AI tools on the job performance and growth of novice programmers (with less than 4 years of experience).
Work & expertiseEmpiricalabstract onlypublisher record
2026-03-01 · Infant and Child Development · Empirical
Chen Cheng
  • Insight. Results showed that starting from age six, children's memory performance improved with the presence of the virtual agent and showed effective use of cognitive offloading through division of labour during encoding.
  • Conclusion. Recognising when to seek external help is essential, especially in the era of advancing AI technology.
Work & expertiseEmpiricalabstract onlypublisher record
2026-05-07 · Center for Open Science · Research article
Arya Jadhav
  • Insight. Drawing from distributed cognition, embodied intelligence, and cognitive offloading theory, we argue that modern AI doesn’t just support mental work—it rewires it.
  • Conclusion. AI tools no longer just assist our thinking—they shape it.
Work & expertiseResearch articleabstract onlypreprint / unreviewed
2026-06-18 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Sudhakar Reddy Gade
  • Insight. The paper grounds it in three bodies of evidence: the established science of cognitive offloading and skill decay, from the Google effect to the measurable shrinkage of spatial memory in habitual users of navigation aids; the classical irony of automation, which leaves the human least practised at exactly the rare task only the human can do; and the first wave of studies on generative AI and critical thinking, which find that confidence in the machine predicts less thinking by the person using it.
  • Conclusion. The first paper in this series argued that leadership work divides into a computable layer that machines absorb and a non-computable layer that they cannot, and it named discernment as the first of four irreducibilities sitting above what it called the Apara Ceiling.
Work & expertiseResearch articleabstract onlypublisher record
2026-07-05 · Center for Open Science · Review
Martina Vacondio, Nicola Vasta, Claudio Mulatti, Francesco Biondi
  • Insight. While performance appears to increase during AI-assisted tasks, participants often show no performance advantage compared to non-users once tools are removed.
  • Conclusion. To navigate these complexities, we outline a future research agenda that emphasizes the need for longitudinal investigations, as well as larger and more diverse participant samples.
Work & expertiseReviewabstract onlypreprint / unreviewed
2026-05-22 · arXiv (Cornell University) · Research article
Sunny Yu, Myra Cheng, Ahmad Jabbar, Ilia Sucholutsky + 3 more
  • Insight. This suggests that completion time itself is not sufficient to characterize efficiency gains.
  • Conclusion. Large language models (LLMs) have the potential to boost human productivity by speeding up task completion -- provided users know when to offload cognitive work to them.
Work & expertiseResearch articlefull textpreprint / unreviewed
2026-06-24 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Misty Richards
  • Insight. As this transition occurs, cognitive offloading increases, reasoning authority is progressively delegated, and human autonomy becomes increasingly vulnerable to erosion.
  • Conclusion. CSA argues that the most consequential governance failures may not emerge through visible malfunction, but through gradual authority migration, where control over framing, salience, interpretation, and reasoning pathways becomes more influential than control over final outputs.
Work & expertiseResearch articleabstract onlypreprint / unreviewed
2026-07-15 · Open MIND · Research article
BrunoKneffel, Fabian Stephany
  • Insight. Large language models (LLMs) can complete knowledge-work tasks on a learner's behalf.
  • Conclusion. Cognitive Load Theory predicts that such cognitive outsourcing raises assisted performance while preventing the schema construction and automation that constitute skill acquisition, whereas a Socratic LLM that withholds answers but scaffolds reasoning should leave learning intact.
Work & expertiseResearch articleabstract onlypreprint / unreviewed
2026-01-01 · SSRN Electronic Journal · Review
Parth Wagh
  • Insight. Frequent or poorly regulated AI use has been associated with cognitive offloading and reduced independent evaluation, while controlled studies of AI-supported learning have also reported improvements in performance and higher-order thinking.
  • Conclusion. The review uses these findings to develop a scaffold–substitute account of AI-assisted cognition and identifies priorities for future research.
Work & expertiseReviewabstract onlypreprint / unreviewed
2026-08-15 · Journal of the Association for Information Systems · Empirical
Haoyue Xu, Yue Xue, Monideepa Tarafdar
  • Insight. We found that increased cognitive offloading can reduce human engagement through uncritical acceptance of GenAI outputs, thereby reducing human compensatory behaviors.
  • Conclusion. However, potential risks in GenAI-generated outcomes, such as hallucinations, can be amplified by cognitive offloading, a process in which humans shift their cognitive effort to GenAI.
Work & expertiseEmpiricalabstract onlypublisher record
2026-04-27 · Formosa Journal of Science and Technology · Research article
Kamaluddin, Waras, Dharma Widada
  • Insight. The results showed that cognitive offloading had a negative effect on the quality of judgment, while metacognition had a positive effect and acted as a partial mediator.
  • Conclusion. The development of artificial intelligence (AI) in organizational decision-making has given rise to the phenomenon of cognitive offloading that has the potential to affect the quality of managerial judgment.
Work & expertiseResearch articleabstract onlypublisher record
2026-04-15 · arXiv (Cornell University) · Empirical
Adriana Caraeni, Alexander Shick, Andrew Lan
  • Insight. In a three-condition sprint-planning experiment, AI-only planning reduced time and cost but missed more risks and produced more rework; human-only planning adapted better but required greater overhead.
  • Conclusion. The authors recommend hybrid planning that delegates estimation and backlog formatting while reserving risk assessment and ambiguity resolution for human deliberation, with longitudinal testing for deskilling.
Work & expertiseEmpiricalfull textpreprint / unreviewed
2026-04-25 · JURNAL PENDIDIKAN MIPA · Review
Tiara Nailah Mahmud, Sri Tiatri, Jap Tji Beng, Vienchenzia Oeyta Dwitama Dinatha + 5 more
  • Insight. Mahasiswa mengandalkan Generative artificial intelligence (GAI) dalam menyelesaikan tugas, dan kegiatan pembelajaran sehari-hari untuk meringankan beban kognitif atau yang disebut cognitive offloading.
  • Conclusion. Penggunaan artificial intelligence dengan model Generative artificial intelligence (GAI) dalam dalam bidang akademik merupakan bentuk inovasi teknologi yang dapat digunakan oleh mahasiswa.
Work & expertiseReviewabstract onlypublisher record
2026-07-30 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Akshat Kandpal, Gurpreet Kaur, Janhvi Gupta, Satyam Kumar Roy + 1 more
  • Insight. Empirical evidence supporting GATO included four mechanisms: decreased ambiguity tolerance, atrophy of structural synthesis processes, contraction of originality, and decreased generative self-efficacy.
  • Conclusion. The use of large language models became widespread to complete high-level generative work including brainstorming, drafting, and constructing arguments.
Work & expertiseResearch articleabstract onlypublisher record
2026-03-23 · arXiv.org · Research article
Kuangzhe Xu, Yu Shen, Longjie Yan, Yinghui Ren
  • Insight. The proliferation of Generative Artificial Intelligence has transformed benign cognitive offloading into a systemic risk of cognitive agency surrender.
  • Conclusion. Driven by the commercial dogma of "zero-friction" design, highly fluent AI interfaces actively exploit human cognitive miserliness, prematurely satisfying the need for cognitive closure and inducing severe automation bias.
Work & expertiseResearch articlefull textpreprint / unreviewed
2026-06-03 · Underline Science Inc. · Empirical
Association for Computational Linguistics 2026, Lei Chen, Shuangyin Li, Jiachuan Wang + 2 more
  • Insight. Our findings demonstrate that BubbleRAG enables LLMs to achieve superior evidence identification abilities typically seen in retrievers, while establishing a cognitive link between external and internal information during answer generation.
  • Conclusion. Retrieval-augmented generation (RAG) extends the capabilities of large language models (LLMs) by providing access to external knowledge.
Work & expertiseEmpiricalfull textpublisher record
2026-03-16 · ACM · Research article
Dániel Szabó, Niels van Berkel, Benjamin Tag, Rune Møberg Jacobsen + 6 more
  • Insight. As a result, there are increasing concerns about the potential deleterious effects of overusing AI in one’s work.
  • Conclusion. Conversational AI now helps people perform a plethora of everyday tasks, especially online.
Work & expertiseResearch articleabstract onlypublisher record
2026-02-09 · arXiv (Cornell University) · Research article
Rose E. Guingrich, Dvija Mehta, Umang Bhatt
  • Insight. People's use of LLM chatbots as thought partners can contribute to cognitive offloading, which can have adverse effects on cognitive skills in cases of over-reliance.
  • Conclusion. Drawing on philosophy, psychology, and computer science research, we clarify the boundary conditions under which belief offloading occurs and provide a descriptive taxonomy of belief offloading and its normative implications.
Work & expertiseResearch articlefull textpreprint / unreviewed
2026-01-21 · arXiv (Cornell University) · Research article
Zhaiyu Fang, Ruipeng Sun
  • Insight. Empirical results demonstrate that AdaTIR reduces tool calls by up to 97.6% on simple tasks and 28.2% on complex challenges while maintaining or enhancing accuracy.
  • Conclusion. Tool-Integrated Reasoning (TIR) has significantly enhanced the capabilities of Large Language Models (LLMs), yet current agents tend to exhibit cognitive offloading, redundantly invoking external tools even for simple tasks.
Work & expertiseResearch articlefull textpreprint / unreviewed
2026-04-02 · Open Access Journal of Economic Research · Conceptual
Vincent English, Brian Kenny
  • Insight. We further propose Organisational Cultural Intelligence (OCQ), an organisation-level dynamic capability, as the key moderator that determines whether cognitive offloading results in beneficial cognitive augmentation or maladaptive cognitive dependence.
  • Conclusion. The paper concludes by discussing implications for theory and practice, offering a new agenda for research and leadership in an era of global, AI-enabled work.
Work & expertiseConceptualabstract onlypublisher record
2026-01-23 · 2026 5th International Symposium on Computer Applications and Information Technology (ISCAIT) · Research article
Qian Peng, Hourong Yu
  • Insight. Evaluations using simulated nighttime care scenarios and expert heuristic assessment show that the assistant reduced the average response time from approximately 17 to 9 minutes, shortened crying episodes, and was associated with lower self reported caregiver anxiety.
  • Conclusion. Nighttime infant crying frequently forces sleep - deprived caregivers to make hasty and error - prone decisions.
Work & expertiseResearch articleabstract onlypublisher record
2025-07-03 · ACM Transactions on Software Engineering and Methodology · Review
Amr Mohamed, Maram Assi, Mariam Guizani
  • Insight. Our findings show that most studies (90%) consider multiple productivity dimensions.
  • Conclusion. We synthesize reported benefits and risks, and apply established conceptual frameworks to map our findings and contextualize their broader implications.
Work & expertiseReviewfull textpreprint / unreviewed
2025-12-01 · Briefings in Bioinformatics · Review
TH Chan, SH Jin, Mun Kay Ho, Joshua W K Ho
  • Insight. Results The framework recommends different degree of LLM usage in different steps in the systematic review process to promote efficiency, ensure accuracy, and prevent cognitive offloading of the researcher involved.
  • Conclusion. Other risks may involve hallucination and responsibility problem.
Work & expertiseReviewabstract onlypublisher record
2025-01-17 · arXiv · Research article
Mahir Akgun, Sacip Toker
  • Insight. The present study contributed to the literature by providing evidence that accessing search tools does not impact individuals’ CSEs in the same way, with some who showed a noticeable change in their CSE after gaining access to search tools, and those who exhibited very little change in their CSE even after they gained access.
  • Conclusion. The search self-efficacy demonstrated here represents one factor that impacts individuals’ CSEs differently.
Work & expertiseResearch articlefull textpreprint / unreviewed
2025-06-26 · arXiv · Research article
Boyu Gou, Zanming Huang, Yuting Ning, Yu Gu + 22 more
  • Insight. In this work, we introduced Mind2Web 2, a novel benchmark specifically designed for compre- hensively evaluating agentic search systems on long-horizon information-gathering tasks and time- varying answers.
  • Conclusion. The views and conclusions contained herein are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the U.S. government.
Work & expertiseResearch articlefull textpreprint / unreviewed
2025-02-12 · arXiv (Cornell University) · Research article
Chenxiang Ma, Zhi‐Sheng Ye, Zhao, Hanyu, Zehua Yang + 8 more
  • Insight. Our evaluation shows that Select-N consistently meets SLOs and improves the serving throughput over existing mechanisms by 1.85X due to maximizing the use of host memory.
  • Conclusion. Offloading large language models (LLMs) state to host memory during inference promises to reduce operational costs by supporting larger models, longer inputs, and larger batch sizes.
Work & expertiseResearch articlefull textpreprint / unreviewed
2025-10-20 · arXiv.org · Empirical
Celeste Riley, Omar Al-Refai, Yadira Colunga Reyes, Eman Hammad
  • Insight. Observations seem to suggest that while AI can substantially enhance memory, creativity, and engagement, it also introduces risks such as diminished critical thinking, skill erosion, and increased anxiety.
  • Conclusion. This paper aims to underscore the need for responsible and context-aware AI design, highlighting gaps for longitudinal research and grounded evaluation frameworks to balance benefits with emerging human-centric risks.
Work & expertiseEmpiricalfull textpreprint / unreviewed
2025-05-20 · arXiv (Cornell University) · Research article
Sebastian Zepf, Mark Colley
  • Insight. Subsequently, we discuss four AI-based superpowers: extended perception, cognitive offloading, externalized memory, and enhanced presence.
  • Conclusion. We then discuss implications for HCI and AI, emphasizing the need for preserving intrinsic human superpowers, identifying meaningful use cases for AI, and evaluating AI's impact on human abilities.
Work & expertiseResearch articlefull textpreprint / unreviewed
2025-12-10 · OSF Preprints (OSF Preprints) · Review
Haneen V A
  • Insight. This PRISMA-registered systematic review maps how Socratic AI and intelligent tutoring designs use questioning, reflection, and cognitive friction to keep learners mentally active.
  • Conclusion. The registry describes the review architecture and intended synthesis; without posted completed results, it cannot establish which intervention design works best.
Work & expertiseReviewabstract onlypreprint / unreviewed
2025-12-03 · Consumer Psychology Review · Research article
Tae Woo Kim, Umair Usman, Aaron Garvey, Adam Duhachek
  • Insight. Through synthesis of the literature on automation, cognitive offloading, and skill acquisition, we demonstrate how the shift from algorithm aversion to AI appreciation creates predictable trajectories in human capability.
  • Conclusion. The paper identifies theoretical mechanisms, boundary conditions, and individual differences that determine whether GenAI use leads to capability enhancement or erosion, proposing testable hypotheses for future empirical investigation.
Work & expertiseResearch articleabstract onlypublisher record
2025-06-15 · Conference on Designing Interactive Systems · Research article
Frederic Gmeiner, Kaitao Luo, Ye Wang, Kenneth Holstein + 1 more
  • Insight. Fundamental cognitive challenges include the need to specify all design criteria as distinct parameters upfront (intent formulation) and designers' reduced cognitive involvement in the design process due to cognitive offloading, which can lead to insufficient problem exploration, underspecification, and limited ability to evaluate outcomes.
  • Conclusion. Based on these findings, we discuss opportunities and tradeoffs of metacognitive support agents and considerations for future AI-based design tools.
Work & expertiseResearch articlefull textpublisher record
2025-09-01 · IEEE Internet Computing · Conceptual
David Andrés García-Barrios
  • Insight. This article proposes and formalizes a multistage conceptual model of the cognitive offloading paradox in human–artificial intelligence (AI) collaboration, identifying when and how intelligent systems augment or deskill human cognition.
  • Conclusion. This framework provides a theoretical foundation for future empirical studies and system development aimed at preserving essential cognitive capacities in increasingly automated environments.
Work & expertiseConceptualabstract onlypublisher record
2025-03-05 · ACM · Research article
Prakash Shukla, Phuong Bui, Sean S Levy, Max Kowalski + 2 more
  • Insight. Our findings indicate that while practitioners express optimism about AI reducing repetitive work and augmenting creativity, they also highlight concerns about over-reliance, cognitive offloading, and the erosion of critical design skills.
  • Conclusion. We argue that UX professionals should critically evaluate AI's role beyond immediate productivity gains and consider its long-term implications for creative autonomy and expertise.
Work & expertiseResearch articlefull textpublisher record
2025-07-02 · arXiv (Cornell University) · Research article
Kenan Tang, Yanhong Li, Yao Qin
  • Insight. Our evaluation of LLM-based components in DIY-MKG shows that vocabulary expansion is reliable and fair across multiple languages, and that the generated quizzes are highly accurate, validating the robustness of DIY-MKG.
  • Conclusion. Existing language learning tools, even those powered by Large Language Models (LLMs), often lack support for polyglot learners to build linguistic connections across vocabularies in multiple languages, provide limited customization for individual learning paces or needs, and suffer from detrimental cognitive offloading.
Work & expertiseResearch articlefull textpreprint / unreviewed
2025-08-22 · Center for Open Science · Empirical
Basil Wahn, Khadija Bawari, Eva Wiese
  • Insight. In conclusion, our results indicate that when it comes to cognitive offloading behavior, a social robot is treated similarly to an algorithm rather than a human interaction partner.
  • Conclusion. Offloading tasks to artificial agents, such as ChatGPT, has become commonplace, yet little research examined offloading of attentionally demanding tasks.
Work & expertiseEmpiricalabstract onlypreprint / unreviewed
2025-11-03 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Bipin, Pearl
  • Insight. This paper outlines a personal information management strategy termed "Cognitive Work Delegation ," which extends the concept of cognitive offloading by leveraging publicly accessible Artificial Intelligence (AI) systems.
  • Conclusion. The core methodology involves the deliberate and open publication of a significant corpus of personal research, essays, and professional documentation.
Work & expertiseResearch articleabstract onlypreprint / unreviewed
2025-04-21 · arXiv (Cornell University) · Empirical
Hongru Wang, Cheng Qian, Wanjun Zhong, Xiusi Chen + 6 more
  • Insight. Experiments with Qwen-2.5 and Qwen-Math across multiple QA benchmarks show that our approach reduces tool calls by up to 68.3\% and improves tool productivity by up to 215.4\%, while maintaining comparable answer accuracy.
  • Conclusion. Tool-integrated reasoning (TIR) augments large language models (LLMs) with the ability to invoke external tools during long-form reasoning, such as search engines and code interpreters, to solve tasks beyond the capabilities of internal reasoning.
Work & expertiseEmpiricalfull textpreprint / unreviewed
2024-11-07 · Alzheimer s & Dementia · Research article
Andrew M. Kiselica, Greta Hermann, Michael K. Scullin, Jared F. Benge
  • Insight. "Everyday technologies" have long been suggested as digital tools to improve life for and enhance care of persons with Alzheimer's disease and related dementias (ADRD).
  • Conclusion. Within this realm, there is a need to balance potential drawbacks of technologies with their ability to positively impact patient and care partner centered outcomes.
Work & expertiseResearch articleabstract onlypublisher record
2024-08-29 · Proceedings of the Human Factors and Ergonomics Society Annual Meeting · Research article
Giovani Diaz Alfaro, Stephen M. Fiore, Kevin Oden
  • Insight. If we are to successfully reduce cognitive workload (CWL), we must adjust how we conceptualize cognition in manned-unmanned teaming environments.
  • Conclusion. Technological advances are increasing the complexity of, interactivity with, and cognitive demands on, human agents, particularly when AI is introduced.
Work & expertiseResearch articleabstract onlypublisher record
2023-10-26 · arXiv (Cornell University) · Research article
Zhicheng Lin
  • Insight. The framework pinpoints both short-term and long-term reasons for engagement, their underlying mechanisms (e.g., cognitive offloading and imaginative stimulation), and the need for a learning mindset to avoid overreliance on AI.
  • Conclusion. Academic writing is an indispensable yet laborious part of the research enterprise.
Work & expertiseResearch articlefull textpreprint / unreviewed
2022-07-14 · Frontiers in Artificial Intelligence · Research article
Sandra Grinschgl, Aljoscha C. Neubauer
  • Insight. In this article, we provide (1) an overview of empirical findings on cognitive offloading and (2) an outlook on how individuals' offloading behavior might change in an AI-enhanced future.
  • Conclusion. Trust in modern AI-technologies might be a strong determinant for wider appropriation and dependence on these technologies to distribute cognition and should thus be considered in an AI-enhanced future.
Work & expertiseResearch articlefull textpublisher record
✚  Health & medicine (10)
2026-02-27 · European Radiology · Research article
Barbara Wichtmann, Daniel Paech, Oleg S. Pianykh, Susie Y. Huang + 3 more
  • Insight. The radiology leadership reply accepts cognitive offloading as a real risk of assistive technology but argues that it should be weighed against patient-care gains rather than used as a categorical reason to reject AI.
  • Conclusion. Its position is governance-oriented: leaders should manage trade-offs, preserve professional autonomy and skills, and judge adoption by patient-centered outcomes.
Health & medicineResearch articlefull textpublisher record
2026-04-30 · International Journal of Innovative Technologies in Social Science · Research article
Jakub Winiarczyk, Karolina Dąbrowska, Emilia Lenkiewicz, Julia Żak + 6 more
  • Insight. However, there are also trade-offs associated with externalizing memory, such as the shift towards "where-to-find" encoding, the potential for altered metacognitive calibration, decreased effort, attentional costs, or dependence on the device.
  • Conclusion. Research objective: This integrative review examines the impact of memory-aided technology on the encoding, storage, retrieval, and coordination of everyday activities and shared information practices: smartphone reminders, cloud storage, wearable lifelogs, clinical cognitive prostheses, and AI personal assistants, as well as the implications that exist at both the individual and institutional levels.
Health & medicineResearch articlefull textpublisher record
2026-06-27 · Zenodo (CERN European Organization for Nuclear Research) · Conceptual
Татяна Йорданова
  • Insight. Instead, it describes a mechanism through which AI-assisted capture and structuring of self-generated thought may reduce the burden on working memory, decrease the fear of losing unfinished ideas, and support the conversion of rapid associative cognition into stable conceptual outputs.
  • Conclusion. It is not a clinical recommendation, diagnosis, therapeutic method, or peer-reviewed medical guideline.
Health & medicineConceptualabstract onlypublisher record
2026-05-05 · arXiv (Cornell University) · Research article
Ilias Chalkidis, Anders Søgaard
  • Insight. The paper finds that cognitive and mental-health risks such as deskilling, critical-thinking atrophy, attachment, and dependence receive little attention in major technical AI-safety reports and venues compared with content and misuse risks.
  • Conclusion. It calls for cognitive and mental-health harms to become explicit AI-safety targets, supported by measurement, public information, and regulation; it is a field-level critique rather than evidence that every GenAI user is harmed.
Health & medicineResearch articlefull textpublisher record
2026-03-19 · ESMO Real World Data and Digital Oncology · Review
Pierre Heudel, Hugo Crochet, Quentin Filori, T. Bachelot + 1 more
  • Insight. Evidence of clinical deskilling, though scarce, is consistent across specialties.
  • Conclusion. Structural deskilling has been reported in cytology following the UK's transition to human papillomavirus primary screening, leading to an 80%-85% reduction in case volumes and consolidation of laboratories from 45 to 8 centers, with major implications for training capacity.
Health & medicineReviewabstract onlypublisher record
2026-01-08 · The British Journal of Psychiatry · Research article
Scott Monteith, Tasha Glenn, John Richard Geddes, Peter C. Whybrow + 3 more
  • Insight. The medical feature article argues that overreliance on AI recommendations may weaken clinical reasoning, decision-making, patient communication, and skill development, especially among younger or less experienced physicians.
  • Conclusion. AI may assist administrative and analytic tasks, but it cannot replace physicians' interpersonal judgment; continuous unaided practice and recognition of model limits remain necessary safeguards.
Health & medicineResearch articlefull textpublisher record
2026-07-01 · Intelligent Hospital · Conceptual
Parvin babaei, Amir Mohammad Chekeni
  • Insight. While discourse on AI in nursing has predominantly centered on patient safety and clinical decision support, a critical dimension remains underexplored: the role of AI-mediated cognitive offloading as a fundamental pillar of professional self-care and workforce sustainability.
  • Conclusion. We argue that the future of nursing self-care hinges not on resilience training or workload reduction alone, but on the deliberate, literate redistribution of cognitive labor between human and machine.
Health & medicineConceptualabstract onlypublisher record
2026-01-21 · arXiv (Cornell University) · Empirical
Matthias Huemmer, Franziska Durner, Theophile Shyiramunda, Michelle J. Cummings-Koether
  • Insight. The ACTIVE Framework synthesizes findings grounded in cognitive load theory: Awareness and task-AI alignment, Critical verification protocols, Transparent human-in-the-loop integration, Iterative skill development countering cognitive offloading, Verification confidence calibration, and Ethical evaluation.
  • Conclusion. This longitudinal pilot study tracked how generative AI reshapes problem-solving over six months across three waves in an academic setting.
Health & medicineEmpiricalfull textpreprint / unreviewed
2025-10-01 · Korean Journal of Medicine · Empirical
Sangzin Ahn
  • Insight. Experimental evidence suggests that while LLM users experience reduced cognitive load across multiple dimensions, their critical reasoning performance may suffer compared with traditional search methods.
  • Conclusion. Rapid advancements in large language models (LLMs) have fundamentally transformed research practices across academic disciplines, with considerable adoption rates among researchers.
Health & medicineEmpiricalfull textpublisher record
2025-11-21 · Frontiers in Psychology · Research article
Runxi Zeng, D. Vance, Jeena Joseph, Ginto Chirayath + 1 more
  • Insight. Evidence suggests that Woebot can reduce symptoms of depression and anxiety in young adults (Li et al., 2025).
  • Conclusion. Its role in mental health is therefore best conceived not in binary terms but as a continuum ranging from healthy offloading to risky overload.At one end of this continuum lies healthy offloading.
Health & medicineResearch articlefull textpublisher record
◎  Cognition & society (35)
Unknown · Semantic Scholar · Research article
Lesia Hrechukha Ph., Olena Pavlenko Ph., Natali і a Butko Ph.
  • Insight. The paper argues that fluent algorithmic agency blurs human–machine boundaries and creates a phenomenological trap: users retain an illusion of control while interpretive judgment and moral responsibility are increasingly externalized.
  • Conclusion. It recommends prospective accountability and friction-by-design so AI systems supplement rather than replace human judgment.
Cognition & societyResearch articleabstract onlypreprint / unreviewed
Unknown · Information Processing & Management · Research article
Heng Ding, Yizhi Shen, Jing Chen, Ping Wang
  • Insight. In a controlled experiment with 130 novices, direct-recommendation AI produced the highest immediate accuracy (68.8%) and fastest decisions, but 7.1% less skill improvement than the no-AI control; moderate and low offloading produced roughly 5.5% skill gains.
  • Conclusion. The skill penalty was linked more to metacognitive miscalibration than low effort, supporting adaptive cognitive friction and competence feedback instead of full automation for novices.
Cognition & societyResearch articleabstract onlypublisher record
Unknown · Semantic Scholar · Research article
Authors not verified in accessible metadata
  • Insight. The paper frames GenAI hallucinations as an assessment problem as well as a technical problem: final products can conceal whether a student reasoned, verified, revised, or merely accepted a fluent error.
  • Conclusion. Its proposed boundary is process-oriented assessment that makes planning, source checking, revision, and attribution of failure visible; the accessible record does not report a controlled outcome study.
Cognition & societyResearch articleabstract onlypreprint / unreviewed
2026-03-04 · Springer Link (Chiba Institute of Technology) · Research article
Wahyu Mahesa Miarta, Muhammad Danu Winata, Rizqi Mutqiyyah, Muhammad Abduh + 1 more
  • Insight. This document investigates the influence of Artificial Intelligence (AI) on communication within Indonesia’s digital society.
  • Conclusion. It systematically reviews the literature and conceptualizes the Prompt Society, which describes the condition wherein communication becomes procedural and algorithmically mediated.
Cognition & societyResearch articleabstract onlypublisher record
2026-07-31 · OpenAlex · Research article
Vrajlal K Sapovadia
  • Insight. This philosophical paper treats curiosity—not information retrieval—as the endangered capacity when people habitually outsource question formation and intellectual exploration to AI.
  • Conclusion. It presents shrinking curiosity as a civilizational risk and argues for preserving active inquiry; the accessible scholarly record does not provide empirical tests for that claim.
Cognition & societyResearch articleabstract onlypublisher record
2026-01-01 · SSRN Electronic Journal · Research article
Akihiro Ito
  • Insight. The theoretical paper defines 'ego decay' as a feedback loop in which sycophantic AI encourages users to stop generating initial judgments and become pass-through nodes, accumulating cognitive debt and weakening metacognition.
  • Conclusion. It proposes a Dynamic Friction System that detects high offloading and injects semantic friction to reactivate user judgment; this is an architectural proposal, not an evaluated intervention.
Cognition & societyResearch articleabstract onlypreprint / unreviewed
2026-08-03 · Computers in Human Behavior · Research article
Kou Tamura, Sayaka Ishibashi, Ahmed Goma, Kenta Yamamoto + 1 more
  • Insight. These findings suggest that LLMs may support cognitive offloading but increase susceptibility among individuals with limited cognitive resources.
  • Conclusion. The ecological generalizability of these findings to everyday dilemma situations remains to be examined in future research.
Cognition & societyResearch articleabstract onlypublisher record
2026 · arXiv.org · Research article
Koutarou Tamura, Sayaka Ishibashi, Ayana Goma, Kenta Yamamoto + 1 more
  • Insight. Across age groups, low decision confidence and high perceived difficulty were associated with greater judgment reversal, suggesting that judgment-specific uncertainty may be more important than general trust in AI in predicting responsiveness to LLM-generated counterarguments.
  • Conclusion. Because trolley dilemmas are abstract philosophical scenarios, future research should examine whether these findings generalize to everyday moral decision-making contexts and to more interactive forms of LLM use.
Cognition & societyResearch articlefull textpreprint / unreviewed
2026-07-16 · Global Alliance of Marketing and Management Associations · Research article
Duan Yihui, Bae Joonheui
  • Insight. The conference paper applies cognitive offloading to AI styling, where delegating outfit selection can reduce decision burden while shifting taste formation and consumer choice toward the system.
  • Conclusion. It frames AI stylists as a convenience–autonomy trade-off; the accessible conference record does not expose enough methods or results for a stronger empirical claim.
Cognition & societyResearch articleabstract onlypublisher record
2026-01-01 · DSpace repository (University of Tartu) · Research article
Annabel Agasild
  • Insight. Käesoleva uurimistöö eesmärgiks oli TI-Hüppe valguses uurida ChatGPT-le liigtuginemistehk valede vastuste omaksvõtmist ja kasutamist, täpsustades tuginemise erinevaid tahke ning seoseid usalduse ja kasutusega.
  • Conclusion. Katsegrupis (n = 325) lahendasid õpilased kriitilist mõtlemist nõudvaid ülesandeid koos ChatGPT pakutud lahendustega, millest ligi pooled (6/11) olid valesti lahendatud.
Cognition & societyResearch articleabstract onlypublisher record
2026-07-20 · arXiv (Cornell University) · Research article
Zixin Chen, Haotian Li, Ziang Xiao, Huamin Qu + 1 more
  • Insight. Across 128,569 naturalistic human–LLM conversations, cognitive engagement appeared in 31.9% of user turns and deeper constructive engagement in 4.9%, showing that everyday use contains selective informal learning as well as answer seeking.
  • Conclusion. AI interactions should be evaluated not only by productivity or displaced effort, but by whether the dialogue preserves opportunities to reason, test ideas, and construct understanding.
Cognition & societyResearch articlefull textpreprint / unreviewed
2026-01-01 · Communications in computer and information science · Research article
Zixin Wang, Silin Liu
  • Insight. Using an eight-code framework, the authors classified 16,851 dialogue turns from 203 undergraduates: 62% reflected passive delegation and 37% active reasoning; more within-student offloading was associated with lower assignment grades.
  • Conclusion. Recovery from offloading back to agentic engagement was the strongest behavioral correlate of performance, shifting attention from whether students use AI to how they use it turn by turn.
Cognition & societyResearch articleabstract onlypublisher record
2026-08-03 · Figshare · Empirical
Kouhei Masumoto
  • Insight. This dataset documents an experiment testing whether LLM-generated counterarguments change moral judgments across younger and older adults, with cognitive functioning, AI trust, and prior LLM experience recorded as possible moderators.
  • Conclusion. It is a reusable data release supporting the associated article, not a separate inferential study; substantive conclusions belong to the linked paper analysis.
Cognition & societyEmpiricalabstract onlypublisher record
2026-01-01 · Open MIND · Conceptual
Strati, Francesco
  • Insight. The formal model identifies a self-reinforcing trap: delegation depreciates autonomous capacity, which expands later delegation into more value-laden decisions, while AI influence can also contaminate the preferences used to evaluate those decisions.
  • Conclusion. It derives domain-contingent friction—minimal for technical tasks and increasing with value-ladenness—as a better policy than uniformly encouraging or discouraging AI use.
Cognition & societyConceptualabstract onlypreprint / unreviewed
2026-03-29 · International Research Journal of Modernization in Engineering Technology and Science · Empirical
Authors not verified in accessible metadata
  • Insight. The survey study asks whether greater AI delegation is associated with weaker self-reported or measured decision-making capacity, treating cognitive offloading as the proposed mechanism.
  • Conclusion. The accessible record does not expose a verifiable sample, analysis, or result, so it supports the research question but not a directional finding.
Cognition & societyEmpiricalabstract onlypublisher record
2026-01-01 · SSRN Electronic Journal · Research article
Yuhe Wei, Yuanyuan Liu
  • Insight. In simulated Level-3 driving, an LLM agent reduced subjective frustration but improved situation awareness only in single-task conditions; dual tasks still produced perception deficits, blink suppression, and heavily occupied visual resources.
  • Conclusion. Trust remained high despite degraded perception, indicating fluent-agent heuristic trust and subjective relief can coexist with objective detachment and risk insensitivity.
Cognition & societyResearch articleabstract onlypreprint / unreviewed
2026-07-03 · Al-Isyrof Jurnal Bimbingan Konseling Islam · Research article
Institut Agama, I. Sunan, Kalijogo Malang, Gatut Setiadi + 2 more
  • Insight. Kemudahan akses terhadap mesin pencari, aplikasi berbasis AI, dan berbagai platform digital mendorong munculnya fenomena cognitive offloading, yaitu kecenderungan mengalihkan sebagian proses kognitif kepada perangkat eksternal.
  • Conclusion. Perkembangan Artificial Intelligence (AI) dan teknologi digital telah mengubah cara remaja memperoleh, mengolah, dan memanfaatkan informasi.
Cognition & societyResearch articlefull textpublisher record
2026-07-08 · MDPI AG · Research article
Chris Fields, Michael Levin
  • Insight. Here we show, employing just basic physics and the Free Energy Principle, that all time-persistent information-processing systems offload information-processing tasks to their environments.
  • Conclusion. We conclude that both theoretical understanding of problem-solving abilities and the engineering of such abilities into artifacts will be improved by considering active computation by the environment as a ubiquitous adjunct to cognition in both living and artificial systems.
Cognition & societyResearch articleabstract onlypreprint / unreviewed
2026-04-22 · Scholars Crossing (Liberty University) · Research article
Ivy Rath
  • Insight. This research seeks to reconcile cognitive offloading with the findings in this new study and demonstrate a tentative cause-and-effect relationship between cognitive offloading and lesser brain activity in generative AI usage.
  • Conclusion. In doing so, generative AI poses unprecedented risk to users' independent cognitive abilities.
Cognition & societyResearch articleabstract onlypublisher record
2026-04-23 · European Archives of Oto-Rhino-Laryngology · Research article
Pasquale Viola, Giuseppe Chiarella
  • Insight. This clinical commentary raises the possibility that routine delegation of attention, memory, and judgment to AI could become a form of cognitive deprivation analogous to underuse of other human capacities.
  • Conclusion. It is a warning and research agenda rather than direct evidence of deprivation; longitudinal clinical measurement is needed before equating AI use with cognitive decline.
Cognition & societyResearch articleabstract onlypublisher record
2026-03-27 · arXiv (Cornell University) · Research article
Tugba Coskun Aslan, Gulser Uncular, Hasan Durmus, Yasin Kavla + 4 more
  • Insight. The findings support the two-factor structure of the scale within the Turkish context, and the instrument demonstrates reliable performance.
  • Conclusion. Wolf, The very efficient assessment of need for cognition: Developing a six-item version.
Cognition & societyResearch articlefull textpreprint / unreviewed
2026-04-13 · CHI Extended Abstracts · Research article
C. Cheatle, Adrian P. Banks
  • Insight. Overall, cognitive offloading impaired confidence calibration, with unaided decision makers showing the greatest alignment between confidence and accuracy.
  • Conclusion. Appropriate AI reliance requires users to accurately evaluate their own performance in order to discern whether to retain or defer responsibility.
Cognition & societyResearch articleabstract onlypublisher record
2025-01-01 · Springer proceedings in business and economics · Research article
Michael Gerlich
  • Insight. Synthesizing two empirical studies, the chapter reports that AI anxieties surface more in anonymous settings and that frequent GenAI reliance correlates with lower structured critical-thinking performance.
  • Conclusion. Its 'trust paradox' explains how people can distrust AI yet depend on it; the chapter recommends environments that preserve critical agency and treat anxiety as diagnostic rather than mere resistance.
Cognition & societyResearch articleabstract onlypublisher record
2025-01-01 · LUTPub (LUT University) · Empirical
Bhattarai, Kushal
  • Insight. The result of reliability test showed that Cronbach’s alpha of variables were above 0.70 ensuring data are reliable for further use and the result of validity test showed that both convergent and discriminant validity were established using Fornell–Larcker criterion.
  • Conclusion. The study investigates the effects of usage of generative AI on critical thinking with cognitive offloading as a mediator.
Cognition & societyEmpiricalfull textpublisher record
2025-09-14 · Jurnal Studi Guru dan Pembelajaran · Research article
Umi Maghfiroh, Ratieh Widhiastuti
  • Insight. Urgensi dari penelitian ini adalah memahami dampak penggunaan kecerdasan buatan (AI) dan cognitive offloading terhadap kemampuan berpikir kritis mahasiswa di era digital dan revolusi industri 4.0, khususnya dalam konteks pendidikan tinggi di Indonesia.
  • Conclusion. Pentingnya urgensi ini diperkuat oleh kebutuhan untuk mengungkap peran Adversity Quotient (AQ) sebagai variabel pemoderasi dalam interaksi kompleks antara teknologi, kognisi, dan kemampuan berpikir kritis.
Cognition & societyResearch articlefull textpublisher record
2025-09-10 · Societies · Research article
Michael Gerlich
  • Insight. This publication is a correction to a table in Gerlich's 2025 Societies article; it does not add a new sample, method, or independent finding about cognitive offloading.
  • Conclusion. Use the corrected article for substantive claims and treat this record only as part of the paper's publication history.
Cognition & societyResearch articleabstract onlypublisher record
2025-01-01 · Acta Psychologica Sinica · Research article
Fancong WANG, Xiaoyu Tang, Shengquan Yu
  • Insight. Graduate students with stronger prior domain knowledge showed more varied interaction sequences—rapid independent planning, precise AI exchanges, selective extraction, and deeper processing—than lower-performing peers.
  • Conclusion. Effective outsourcing was not passive delegation: it depended on active participation and balanced, tightly connected internal and external cognitive networks.
Cognition & societyResearch articleabstract onlypublisher record
2025-12-18 · Zenodo (CERN European Organization for Nuclear Research) · Research article
Prasoon Jadon
  • Insight. While AI improves efficiency and accessibility, emerging research suggests that excessive reliance on AI may reduce active cognitive engagement.
  • Conclusion. Using secondary data sources and conceptual analysis, this study argues for balanced AI integration that augments rather than replaces human cognition.
Cognition & societyResearch articleabstract onlypreprint / unreviewed
2025-08-01 · Cognitive Science · Research article
Kristy L. Armitage, Alicia K. Jones, Jonathan Redshaw
  • Insight. These results provide preliminary evidence that pre‐exposure to indiscriminate cognitive offloading can promote selectivity in children's subsequent strategy use, though this selectivity may reflect a desire to avoid cognitive effort rather than improve task performance.
  • Conclusion. Abstract With the rise of wearable technologies, mobile devices and artificial intelligence comes a growing pressure to understand downstream effects of cognitive offloading on children's future thinking and behavior.
Cognition & societyResearch articleabstract onlypublisher record
2025-08-21 · Minds and Machines · Research article
Avigail Ferdman
  • Insight. The paper argues that AI-mediated democratic deliberation can deskill capacities needed for public reason—including patience, moral attention, justice, joint action, and reciprocal justification—by reducing embodied shared practice.
  • Conclusion. AI may improve scale and policy understanding, but deliberative systems should combine digital and non-digital participation so efficiency does not erode civic skills and mutual assurance.
Cognition & societyResearch articlefull textpublisher record
2025-11-05 · AI & Society · Research article
Avigail Ferdman
  • Insight. Abstract Many artificial intelligence tools replace or stand to replace human activity, via automated decision-making, recommender systems and content generation.
  • Conclusion. This paper argues for applying a structural perspective to this phenomenon.
Cognition & societyResearch articlefull textpublisher record
2024-12-04 · Social Behavior and Personality An International Journal · Empirical
Tao Wei, Mengqiu Zhang, Yi-Chuan Liu
  • Insight. Although large language models like ChatGPT enable creators to outsource cognitive tasks, partially or entirely entrusting the creative process to artificial intelligence (AI), there is no standardized tool for measuring cognitive outsourcing.
  • Conclusion. Our research underscores the significance of cognitive outsourcing in AI-enhanced creativity, providing a comprehensive tool for future investigations into the effects of AI on human creativity and practices.
Cognition & societyEmpiricalabstract onlypublisher record
2024 · LAK Workshops · Research article
S. B. Shum
  • Insight. Through examples in argument analysis and summarizing learning outcomes, the paper shows that GenAI can perform tasks previously treated as distinctly demanding forms of critical analysis.
  • Conclusion. The author argues that whether this becomes augmentation or agency loss depends on tool design and use: AI should provoke questioning and critical evaluation rather than silently replace them.
Cognition & societyResearch articleabstract onlypreprint / unreviewed
2024 · Elsevier BV · Research article
Michael Gerlich
  • Insight. The study links frequent AI-tool use and higher trust in AI with greater cognitive offloading and lower critical-thinking performance, framing trust as a mechanism that can make delegation feel safe.
  • Conclusion. The evidence is correlational rather than proof of atrophy; it supports calibrated trust, critical-thinking practice, and designs that require users to inspect AI outputs.
Cognition & societyResearch articleabstract onlypreprint / unreviewed
2023-12-01 · Iowa State University · Research article
Karthika Suresh Kumar
  • Insight. This student research project examines whether using ChatGPT as an external memory aid changes retention, explicitly treating the tool as a cognitive-offloading intervention.
  • Conclusion. The accessible repository metadata establishes the topic but does not expose enough methods or results to claim that ChatGPT improved or impaired memory retention.
Cognition & societyResearch articleabstract onlypublisher record