The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers
Open accessThe DOI matches its Crossref record; 132 of the 132 references it cites were found in a registry or an index; 23 of the 39 sources its claims cite were checked for support.
Abstract
The rise of Generative AI (GenAI) in knowledge workflows raises questions about its impact on critical thinking skills and practices.We survey 319 knowledge workers to investigate 1) when and how they perceive the enaction of critical thinking when using GenAI, and 2) when and why GenAI affects their effort to do so.Participants shared 936 first-hand examples of using GenAI in work tasks.Quantitatively, when considering both task-and user-specific factors, a user's task-specific self-confidence and confidence in GenAI are predictive of whether critical thinking is enacted and the effort of doing so in GenAI-assisted tasks.Specifically, higher confidence in GenAI is associated with less critical thinking, while higher self-confidence is associated with more critical thinking.Qualitatively, GenAI shifts the nature of critical thinking toward information verification, response integration, and task stewardship.Our insights reveal new design challenges and opportunities for developing GenAI tools for knowledge work.
Overview
CoChat Summary
TL;DR
A survey of 319 knowledge workers who use GenAI at work at least weekly, drawing on 936 real-world examples, finds that higher confidence in GenAI is associated with less critical thinking, while higher self-confidence is associated with more. GenAI also shifts perceived critical-thinking effort toward information verification, response integration, and task stewardship.
Key findings
- Participants reported enacting critical thinking in 555 of 936 examples (59.29%). Higher confidence in AI doing the task negatively correlated with perceived enaction of critical thinking (β = -0.69, p < 0.001); confidence in doing the task oneself (β = 0.26, p = 0.026), confidence in evaluating AI responses (β = 0.31, p = 0.046), and overall tendency to reflect on work (β = 0.52, p < 0.001) positively correlated with perceived enaction.
- Perceived effort decreased in the majority of examples: “much less effort” or “less effort” comprised 72% for Knowledge, 79% for Comprehension, 69% for Application, 72% for Analysis, 76% for Synthesis, and 55% for Evaluation. Higher confidence in AI negatively correlated with perceived effort for five of six activities (all except Application): Knowledge (β = -0.11, p = 0.029), Comprehension (β = -0.13, p = 0.014), Analysis (β = -0.15, p = 0.003), Synthesis (β = -0.12, p = 0.026), and Evaluation (β = -0.23, p < 0.001).
- Qualitatively, GenAI shifted critical-thinking effort in three ways: for Knowledge and Comprehension, from information gathering to information verification; for Application, from problem-solving to AI response integration; and for Analysis, Synthesis, and Evaluation, from task execution to task stewardship.
- Motivators for critical thinking included work quality (74/319), potential negative outcomes (116/319), and skill development (13/319). Inhibitors included awareness barriers such as secondary use of GenAI (14/319), trivial tasks (55/319), and trust/reliance on GenAI (83/319); motivation barriers such as lack of time (44/319) and not being part of job responsibilities (11/319); and ability barriers such as barriers to inspecting AI responses (58/319) and revising queries/improving responses (72/319).
Methods
- Online survey on Prolific of 319 knowledge workers who self-reported using GenAI tools at work at least once per week. The study received 333 responses and excluded 14 for low response quality; retained participants spent an average of 43.19 minutes (SD = 23.13) completing the survey and were compensated GBP £10.
- Participants shared three real-world GenAI work examples each. The study collected 957 examples, removed 11 lacking sufficient detail and 11 duplicated or non-GenAI examples, retaining 936: 374 Creation (39.96%), 303 Information (32.37%), and 259 Advice (27.67%).
- Quantitative analysis used one random-intercepts logistic regression for perceived enaction of critical thinking and six random-intercepts linear regressions for perceived effort across Bloom’s taxonomy activities, with Participant ID as a random intercept. Benjamini-Hochberg correction was applied across 98 hypothesised predictors, yielding a corrected p-value threshold of 0.007. Qualitative analysis used open coding of free-text responses; one researcher initially coded 50 responses with three others to build a codebook, then two researchers coded the remaining 269 responses.
Limitations
- The paper states that participants occasionally conflated reduced effort in using GenAI with reduced effort in critical thinking with GenAI, potentially leading to inaccurate self-reporting; future studies could use think-aloud protocols or task-based assessments.
- The paper states that subjective task confidence may not be well-calibrated with objective expertise, and that the survey was English-only, with no representation of non-English-speaking or multilingual contexts.
- The paper states that the sample was biased toward younger, more technologically skilled participants who use GenAI at work at least weekly, and that GenAI tools are constantly evolving; the task taxonomy was coarse-grained, and future longitudinal or telemetry/experience-sampling work is suggested. The text also makes evident that the analysis does not establish causation, as the paper itself notes.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures
Verification
Claims the paper makes
12 Valid · 5 Not valid · 9 Not checked · 146 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“These findings suggest that a reflective approach toward the use of GenAI tools, which can lead to what prior work refers to as”
§ The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers > 4.2.1 Task Factors.
What the sources say
- Partly supportedcites [20] 20. Valerie Chen, Q. Vera Liao, Jennifer Wortman Vaughan, and Gagan Bansal. 2023. Understanding the Role of Human Intuition on Reliance in Human- AI Decision-Making with Explanations. Proceedings of the ACM on Human- Computer Interaction 7, CSCW2 (Sept. 2023), 1-32. https://doi.org/10.1145/ 3610219 sourceExperimental — accuracy not yet measured
Section 4.1 (page unavailable) reports that participants with strong intuition sometimes discounted an AI prediction, supporting a link between confidence in their own judgment and critical use of AI. But it also says they sometimes decided quickly without checking explanations when the AI agreed with them. The study concerns two AI-assisted prediction tasks, not GenAI use by knowledge workers, and does not establish the claim’s general likelihood comparison. (LLM, conf=0.86)
hedgedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on 2 assumptions not shown here.
- Partly supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Quantitatively, when considering both task-and user-specific factors, a user's task-specific self-confidence and confidence in GenAI are predictive of whether critical thinking is enacted and the effort of doing so in GenAI-assisted tasks.”
§ The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers > Abstract
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Specifically, higher confidence in GenAI is associated with less critical thinking, while higher self-confidence is associated with more critical thinking.”
§ The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers > Abstract
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“The data shows a shift in cognitive effort as knowledge workers increasingly move from task execution to oversight when using GenAI.”
§ The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers > 1 Introduction
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the study reveals that knowledge workers engage in critical thinking when using GenAI tools primarily to ensure the quality of their work.”
§ The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers > 1 Introduction
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Rests on an assumption not shown here.
78 more claims
Create a free account to see all 172 claims and to verify the 1 reference still open.
- Probably not a substantial claim. 68
- It may state two separate claims as one. 35
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“our findings show that GenAI may undermine students’ motivation to engage in the cognitive effort that is necessary for deeper intellectual enjoyment (contra Lee et al., 2025; see also Stadler et al., 2024).”
From If ChatGPT can do it, where is my creativity? generative AI boosts performance but diminishes experience in creative writing (2025) · cites this paper as Lee et al., 2025
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“We might also be concerned about mechanised convergence in the increased use of GenAI; that AI generation tends to result in less diverse textual output (Lee et al., 2025).”
From If ChatGPT can do it, where is my creativity? generative AI boosts performance but diminishes experience in creative writing (2025) · cites this paper as Lee et al., 2025
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Further, GenAI tools can increase developers' perceived productivity [Ziegler et al. 2024] and level of understanding [Prather et al. 2024] without delivering actual improvements, potentially leading to a false sense of confidence [Lee et al. 2025] among novices.”
From How Do Programming Students Use Generative AI? (2025) · cites this paper as Lee et al. 2025
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Lee et al. [8] found that professionals using GenAI tools reported reduced cognitive effort but increased confidence in the quality of their outputs, revealing a mismatch between perceived and actual engagement.”
From From Offloading to Engagement: An Experimental Study on Structured Prompting and Critical Reasoning with Generative AI (2025) · cites this paper as [8]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“In the present study, this same pattern was evident in the unguided condition. Although participants often believed they had 'used AI to check their ideas,' interview data revealed that many had in fact accepted the AI's structure wholesale.”
From From Offloading to Engagement: An Experimental Study on Structured Prompting and Critical Reasoning with Generative AI (2025) · cites this paper as Lee et al. [8]
Experimental — accuracy not yet measured
20 more claims about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing
Benjamini, Y., & Hochberg, Y. (1995) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1995.tb02031.x · Cited by 109,482 · cited 1×
“To correct for multiple comparisons, we apply the Benjamini–Hochberg procedure [9] with a total of 98 hypothesised predictors across the seven models, yielding a corrected p-value threshold of 0.007.”
InfluentialBasics of Qualitative Research
Ozanne, J. L., Strauss, A., & Corbin, J. (1992) · Journal of Marketing Research · doi:10.2307/3172751 · Cited by 17,850
Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019) · International Journal of Educational Technology in Higher Education · doi:10.1186/s41239-019-0171-0 · Cited by 7,242 · cited 1×
“To mitigate this, using GenAI for individualised, content-focused feedback may help novice writers develop writing skills while improving productivity [58, 86, 144].”
Binder, J. R., Desai, R. H., Graves, W. W., & Conant, L. L. (2009) · Cerebral Cortex · doi:10.1093/cercor/bhp055 · Cited by 4,060 · cited 1×
“Cognitive science shows that effective learning requires “grounding” information through multiple perspectives and examples [10, 11, 68], and GenAI can offer personalised analogies to aid this process [77, 90].”
Dewey, J. (1910) · D C Heath eBooks · doi:10.1037/10903-000 · Cited by 3,998
Human trust in artificial intelligence: Review of empirical research
Glikson, E., & Woolley, A. W. (2020) · Academy of Management Annals · doi:10.5465/annals.2018.0057 · Cited by 2,921 · cited 1×
“We measure participants’ overall trust in GenAI, which has been shown to correlate with users’ attitudes and adoption of the use of the technologies [43, 76].”
Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence
Noy, S., & Zhang, W. (2023) · doi:10.2139/ssrn.4375283 · Cited by 2,157
The neurobiology of semantic memory
Binder, J. R., & Desai, R. H. (2011) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2011.10.001 · Cited by 2,114 · cited 1×
“Cognitive science shows that effective learning requires “grounding” information through multiple perspectives and examples [10, 11, 68], and GenAI can offer personalised analogies to aid this process [77, 90].”
Principles of mixed-initiative user interfaces
Horvitz, E. (1999) · Proceedings of the SIGCHI conference on Human factors in computing systems the CHI is the limit - CHI '99 · doi:10.1145/302979.303030 · Cited by 1,490 · cited 1×
“Proactive systems take the initiative [52] to interrupt the user to highlight the need and opportunity for critical thinking in situations where it is likely to be overlooked; a reactive approach would allow the user to explicitly request critical thinking assistance when it is consciously needed.”
Generative artificial intelligence enhances creativity but reduces the diversity of novel content
Doshi, A., & Hauser, O. (2023) · doi:10.2139/ssrn.4535536 · Cited by 1,350
Research works printed but not linked: 19
- 12. Benjamin S Bloom, Max D Engelhart, Edward J Furst, Walquer H Hill, David R Krathwohl, et al. 1956. Taxonomy of educational objetives: the classification of educational goals: handbook I: cognitive domain. Technical Report. New York, US: D. Mckay.
- 19. Michael Castelluccio. 2022. IS DIGITAL AMNESIA REAL? Strategic Finance 104, 3 (2022), 57-58.
- 21. Ooi Yan Chiew, An Qi Lai, and Wen Xin Liew. 2020. Digital technology overuse as a predictor of digital amnesia and productivity. Ph. D. Dissertation. UTAR.
- 27. Amir Dirin, Ari Alamäki, and Jyrki Suomala. 2019. Digital amnesia and personal dependency in smart devices: A challenge for AI. Proceedings of Fake Intelligence Online Summit 2019 (2019).
- 37. Peter A Facione et al. 2011. Critical thinking: What it is and why it counts. Insight assessment 1, 1 (2011), 1-23.
- 38. Peter A Facione, Carol A Sanchez, Noreen C Facione, and Joanne Gainen. 1995. The disposition toward critical thinking. The Journal of General Education 44, 1 (1995), 1-25.
- 40. Mary Forehand. 2010. Bloom's taxonomy. Emerging perspectives on learning, teaching, and technology 41, 4 (2010), 47-56.
- 54. William Huitt. 2011. Bloom et al. 's taxonomy of the cognitive domain. Educa- tional psychology interactive 22 (2011), 1-4.
- 58. Hong Jiao and Robert W Lissitz. 2020. Application of Artificial Intelligence to Assessment. IAP.
- 70. Patricia M King. 1997. The reflective judgment model: Transforming assump- tions about knowing. College student development and academic life: psychologi- cal, intellectual, social, and moral issues 4 (1997), 141.
- 74. Deanna Kuhn. 1993. Connecting scientific and informal reasoning. Merrill- Palmer Quarterly (1982-) (1993), 74-103.
- 77. George Lakoff and Mark Johnson. 2008. Metaphors we live by. University of Chicago press.
- 86. Jill Burstein McCaffrey, Brian Riordan, and Daniel. 2020. Expanding Automated Writing Evaluation. In Handbook of Automated Scoring. Chapman and Hall/CRC.
- 92. Richard L Miller and William Wozniak. 2001. Counter-attitudinal advocacy: Effort vs. self-generation of arguments. Current Research in Social Psychology 6, 4 (2001), 46-55.
- 102. Samir Passi and Mihaela Vorvoreanu. 2022. Overreliance on AI literature review. Microsoft Research (2022).
- 103. Richard Paul and Linda Elder. 2020. The miniature guide to critical thinking concepts and tools (8th edition ed.). Rowman & Littlefield, Lanham, Md. OCLC: on1132213785.
- 108. Sebastian Raisch and Kateryna Fomina. 2023. Combining human and artificial intelligence: Hybrid problem-solving in organizations. Academy of Management Review ja (2023), amr-2021.
- 118. Advait Sarkar. 2024. Large Language Models Cannot Explain Themselves. In ACM CHI 2024 Workshop on Human-Centered Explainable AI (HCXAI).
- 139. Donna Wilson and Marcus Conyers. 2016. Teaching students to drive their brains: Metacognitive strategies, activities, and lesson ideas. Ascd.
Cited by
Tian, J., & Zhang, R. (2025) · Acta Psychologica · doi:10.1016/j.actpsy.2025.105725 · Cited by 232
Liu, X., & Zhong, B. (2025) · Educational Research Review · doi:10.1016/j.edurev.2025.100741 · Cited by 78
Rodger, D., Mann, S. P., Earp, B., Savulescu, J., Bobier, C., & Blackshaw, B. P. (2025) · Nurse Education in Practice · doi:10.1016/j.nepr.2025.104461 · Cited by 64
“Furthermore, there is evidence from other domains showing that the use of GenAI can diminish critical thinking and independent problem-solving skills over time (Lee et al., 2025).”
Mei, P., Brewis, D. N., Nwaiwu, F., Sumanathilaka, D., Alva-Manchego, F., & Demaree-Cotton, J. (2025) · Computers in Human Behavior: Artificial Humans · doi:10.1016/j.chbah.2025.100140 · Cited by 50
AI deskilling is a structural problem
Ferdman, A. (2025) · AI & Society · doi:10.1007/s00146-025-02686-z · Cited by 108
Rana, V., Verhoeven, B., & Sharma, M. (2025) · Journal of University Teaching and Learning Practice · doi:10.53761/tjse2f36 · Cited by 48
Gerlich, M. (2025) · Data · doi:10.3390/data10110172 · Cited by 40
“Such patterns can erode the perceived value of human input, an observation echoed in the findings of Lee et al. [8], who noted that many workers reported growing reliance on AI tools while showing decreased confidence in their own analytical contributions.”
InfluentialHow Do Programming Students Use Generative AI?
Rahe, C., & Maalej, W. (2025) · Proceedings of the ACM on Software Engineering · doi:10.1145/3715762 · Cited by 45
“Further, GenAI tools can increase developers’ perceived productivity [Ziegler et al. 2024] and level of understanding [Prather et al. 2024] without delivering actual improvements, potentially leading to a false sense of confidence [Lee et al. 2025] among novices.”
Beyond assistance: embracing AI as a collaborative co-agent in education
Katsenou, R., Kotsidis, K., Papadopoulou, A., Anastasiadis, P., & Deliyannis, I. (2025) · Education Sciences · doi:10.3390/educsci15081006 · Cited by 43
Biases in AI: acknowledging and addressing the inevitable ethical issues
Hofmann, B. (2025) · Frontiers in Digital Health · doi:10.3389/fdgth.2025.1614105 · Cited by 62
“Additionally, we need to pay special attention to Residual bias after mitigating various types of bias in AI. biases generated by AI, such as overreliance (162), deskilling (77, 78), and acceptance of algorithmic discrimination (162) as they can proliferate or enhance existing bias.”
Yankouskaya, A., Babiker, A., Rizvi, S., Alshakhsi, S., Liebherr, M., & Ali, R. (2025) · ACM Transactions on the Web · doi:10.1145/3765895 · Cited by 32
“Research suggests lower self-confidence could reduce an individual ’s critical thinking and cognitive effort, which may further increase dependency on these systems, possibly resulting in uncritical acceptance of the information provided to them [42].”
Vendrell, M., & Johnston, S.-K. (2026) · Computers and Education: Artificial Intelligence · doi:10.1016/j.caeai.2026.100572 · Cited by 43
“Similarly, Lee et al. (2025) found that confidence in AI tools was associated with lower levels of critical engagement, while self-confidence predicted more thoughtful reasoning.”
The Memory Paradox: Why Our Brains Need Knowledge in an Age of AI
Oakley, B., Johnston, M., Chen, K., Jung, E., & Sejnowski, T. (2025) · doi:10.2139/ssrn.5250447 · Cited by 31
ChatGPT as a cognitive crutch: Evidence from a randomized controlled trial on knowledge retention
Barcaui, A. (2025) · Social Sciences & Humanities Open · doi:10.1016/j.ssaho.2025.102287 · Cited by 51
Elshall, A. S., & Badir, A. (2025) · Frontiers in Education · doi:10.3389/feduc.2025.1596462 · Cited by 28
Generative artificial intelligence in education: (what) are we thinking?
Costa, C., & Murphy, M. G. (2025) · Learning Media and Technology · doi:10.1080/17439884.2025.2518258 · Cited by 59
Quality work in the future: New directions via a co-evolving sociotechnical systems perspective
Parker, S. K., Ballard, T., Billinghurst, M., Collins, C. G., Dollard, M. F., Griffin, M. A., Johal, W., Jorritsma, K., Kowalkiewicz, M., Kyndt, E., et al. (2025) · Australian Journal of Management · doi:10.1177/03128962251331813 · Cited by 52
“A recent study (Lee et al., 2025) showed that when people use generative AI systems to improve their efficiency, people with high confidence in the AI system engage less in critical thinking, while those with high confidence in their own abilities engage more in critical thinking.”
The cognitive mirror: a framework for AI-powered metacognition and self-regulated learning
Tomisu, H., Ueda, J., & Yamanaka, T. (2025) · Frontiers in Education · doi:10.3389/feduc.2025.1697554 · Cited by 53
ChatGPT is incredible (at being average)
Rudko, I., & Bonab, A. B. (2025) · Ethics and Information Technology · doi:10.1007/s10676-025-09845-2 · Cited by 32
Palma, G., Cecchi, G., & Rizzo, A. (2025) · Electronics · doi:10.3390/electronics14102061 · Cited by 37
“Lee et al. [21] examined the cognitive effects of generative AI, specifically focusing on how LLMs impact human critical thinking and cognitive effort in predictive maintenance decision-making within knowledge work contexts by surveying knowledge workers.”
Huang, S., Long, L., Zhu, Y., & Zhu, J. N. Y. (2025) · International Journal of Information Management · doi:10.1016/j.ijinfomgt.2025.102986 · Cited by 28
Artificial intelligence agents in cancer research and oncology
Truhn, D., Azizi, S., Zou, J., Cerdá-Alberich, L., Mahmood, F., & Kather, J. N. (2026) · Nature reviews. Cancer · doi:10.1038/s41568-025-00900-0 · Cited by 46
Leveraging large language models for thematic analysis: a case study in the charity sector
Wen, C., Clough, P., Paton, R., & Middleton, R. (2025) · AI & Society · doi:10.1007/s00146-025-02487-4 · Cited by 45
“The relative ease with which LLMs can be used for tasks, such as thematic analysis, can also result in an unhealthy trust in the quality of outputs resulting in a lack of critical analysis and engagement with the process (Lee et al. 2025).”
Agentic AI and the rise of in silico team science in biomedical research
Li, B., Saini, A. K., Hernandez, J. G., & Moore, J. H. (2026) · Nature Biotechnology · doi:10.1038/s41587-026-03035-1 · Cited by 45
“For example, individual users tend to use less critical thinking (for example, recalling, combining and judging ideas) when using generative AI tools, particularly when they have more confidence in the tools than in their own ability to solve the tasks at hand 149 .”
Oligarchy, State, and CryptopiaPreprint
Cohen, J. E. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5171050 · Cited by 28
Teaching with the algorithm: anticipations of AI in teacher education
Stenberg, C.-J., Åkerfeldt, A., Stenliden, L., & McGrath, C. (2025) · European Journal of Teacher Education · doi:10.1080/02619768.2025.2540793 · Cited by 27
“In Zuboff’s work, this resulted in workers either ‘giving in’ to the systems by resigning professional agency and responsibility, or workers engaging actively with the output and understanding its relation to their work practices (Lee et al. 2025).”
Dutta, D., & Naveen, P. M. (2025) · Human Resource Management · doi:10.1002/hrm.70018 · Cited by 34
Generative Artificial Intelligence and the Future of Public Knowledge
Spennemann, D. H. R. (2025) · Knowledge · doi:10.3390/knowledge5030020 · Cited by 18
“…[137] and an attitude that ‘near enough is good enough,’ the bulk of the public will avail themselves of solutions that provide the immediate and most convenient answers with the least amount of effort (cognitive offloading)—especially where confidence in the abilities of AI is high [142,143].”
Vicent, M., Katushabe, C., Gloria, M., & Rugasira, J. (2025) · Discover Artificial Intelligence · doi:10.1007/s44163-025-00495-3 · Cited by 40
Explainable AI-Driven Quality and Condition Monitoring in Smart Manufacturing
Ahangar, M. N., Farhat, Z. A., Sivanathan, A., Ketheesram, N., & Kaur, S. (2026) · Sensors · doi:10.3390/s26030911 · Cited by 41
Liang, Z., Du, J., & Hua, Y. (2025) · Journal of theoretical and applied electronic commerce research · doi:10.3390/jtaer20030236 · Cited by 43
“The emergence of the Internet, along with a wave of new digital technologies—such as information and communication technology (ICT), artificial intelligence (AI) [1,2], cloud computing, big data—is profoundly transforming the global economy and society [3].”
Augmenting diversity in hiring decisions with artificial intelligence tools
Wilkens, U., Lutzeyer, I., Zheng, C. S.-M., Beşer, A., & Prilla, M. (2025) · The International Journal of Human Resource Management · doi:10.1080/09585192.2025.2492867 · Cited by 24
“Moreover, intensive tool use reduces decision makers' cognitive effort on own ideas (Lee et al., 2025).”
InfluentialWhen Autonomy Breaks: The Hidden Existential Risk of AI
Krook, J. (2025) · AI & SOCIETY · doi:10.1007/s00146-025-02397-5 · Cited by 14
What Needs Attention? Prioritizing Drivers of Developers' Trust and Adoption of Generative AI
Choudhuri, R., Trinkenreich, B., Pandita, R., Kalliamvakou, E., Steinmacher, I., Gerosa, M., Sanchez, C., & Sarma, A. (2025) · arXiv · doi:10.48550/arxiv.2505.17418 · Cited by 14
Abbas, M., Khan, T. I., & Jam, F. A. (2025) · Journal of Academic Ethics · doi:10.1007/s10805-025-09659-3 · Cited by 63
“Nonetheless, the excessive use of this tool may yield potential hazards associated with it (Abbas et al., 2024 ; Lee et al., 2025).”
AI Could Have Written This: Birth of a Classist Slur in Knowledge Work
Sarkar, A. (2025) · CHI Conference Extended Abstracts on Human Factors in Computing Systems (CHI EA) · doi:10.1145/3706599.3716239 · Cited by 27
Self-citationOng, A. Y., Merle, D. A., Wagner, S. K., & Keane, P. A. (2025) · Journal of Medical Internet Research · doi:10.2196/75666 · Cited by 14
AI-overdependence and human cognitive decline: Hazards, evidence, and mitigation strategies
Noorbehbahani, F., & Oyibo, K. (2026) · Computers in Human Behavior Reports · doi:10.1016/j.chbr.2026.101102 · Cited by 12
Melanou, C., & Beege, M. (2026) · Education Sciences · doi:10.3390/educsci16040651 · Cited by 12
Building organizational and project readiness for AI: A dialectical approach
Armanios, D. E., & Tucci, C. L. (2025) · International Journal of Project Management · doi:10.1016/j.ijproman.2025.102792 · Cited by 22
When humans and large language models collaborate, problem-finding illuminates
Sabbah, J., & Li, F. (2025) · Innovation · doi:10.1080/14479338.2025.2504428 · Cited by 21
“Similarly, Lee et al. (2025) find that among knowledge workers using GenAI, greater confidence in GenAI correlates with reduced critical thinking, whereas higher self-confidence is linked to enhanced critical thinking.”
Vibe coding: programming through conversation with artificial intelligence
Sarkar, A., & Drosos, I. (2025) · arXiv · doi:10.48550/arxiv.2506.23253 · Cited by 11
“This level of expertise is less about writing code line-by-line (as might be accelerated by earlier tools) and more about orchestrating the model and other tools within a complex development environment (what Lee et al. (2025) term “task stewardship”).”
Self-citation“Who will be left behind?”: A Swedish case of learning AI in vocational education
Metreveli, A., Chen, X., Hedman, A., & Sergeeva, A. (2025) · International Journal of Educational Research · doi:10.1016/j.ijer.2025.102697 · Cited by 24
Kreijkes, P., Kewenig, V. N., Kuvalja, M., Lee, M., Hofman, J. M., Vitello, S., Sellen, A. J., Rintel, S., Goldstein, D. G., Rothschild, D. M., et al. (2025) · Computers & Education · doi:10.1016/j.compedu.2025.105514 · Cited by 51
Self-citationThe Impact of Digital Safety Competence on Cognitive Competence, AI Self-Efficacy, and Character
Ma, C. M. S., Shek, D. T. L., Fan, I. Y. H., Zhu, X., & Hu, X. (2025) · Applied Sciences · doi:10.3390/app15105440 · Cited by 24
Fostering programming skill and critical thinking through AI-assisted PBL integration
Omeh, C. B., Ayanwale, M. A., Mnguni, L. E., & Olelewe, C. J. (2025) · Journal of New Approaches in Educational Research · doi:10.1007/s44322-025-00041-0 · Cited by 33
“This finding is consistent with Lee et al (2025) who demonstrated that the impact of Gen AI in critical thinking reduces cognitive effort.”
The Epistemic Downside of Using LLM-Based Generative AI in Academic Writing
Tang, B. L. (2025) · Publications · doi:10.3390/publications13040063 · Cited by 24
“Another survey by Lee and colleagues found that higher confidence in generative AI is associated with less critical thinking, while higher self-confidence is associated with more critical thinking (Lee et al., 2025).”
How AI is rewiring the human brain: the generational transformation of cognition and knowing
Westerbeek, H. M. (2026) · AI & Society · doi:10.1007/s00146-026-02912-2 · Cited by 19
“Lee et al. (2025) draw attention to the growing gulf between productivity and retention, workers who complete more tasks, but understand and remember less.”
Chat-IRB? How application-specific language models can enhance research ethics review
Mann, S. P., Seah, J. J., Latham, S. R., Savulescu, J., Aboy, M., & Earp, B. D. (2025) · Journal of Medical Ethics · doi:10.1136/jme-2025-110845 · Cited by 17
“38 Moreover, any perception of IRB over-reliance on or deference to LLMs would challenge existing societal expectations that ethical decision-making should be performed exclusively by human persons.”
Garcia Ramos, J. (2025) · Frontiers in Education · doi:10.3389/feduc.2025.1654805 · Cited by 8
Generative AI, ESG Sensemaking, and Environmental Performance: an OIPT Perspective
Bag, S., Srivastava, G., Routray, S., & Chiarini, A. (2026) · Business Strategy and the Environment · doi:10.1002/bse.70520 · Cited by 19
The brain side of human-AI interactions in the long-term: the “3R principle”
Rossi, S., Fraccaro, V., & Manzotti, R. (2026) · npj Artificial Intelligence · doi:10.1038/s44387-025-00063-1 · Cited by 15
van Zyl, L. E. (2026) · The Journal of Positive Psychology · doi:10.1080/17439760.2026.2632939 · Cited by 9
GhostCite: A Large-Scale Analysis of Citation Validity in the Age of Large Language Models
Xu, Z., Qiu, Y., Sun, L., Miao, F., Wu, F., Li, X., Wang, X., Lu, H., Zhang, Z., Hu, Y., et al. (2026) · arXiv · doi:10.48550/arxiv.2602.06718 · Cited by 3
“But whether researchers’ self-reported verification behaviors match their actual practices has not been systematically studied: it remains unclear how often researchers rely on AI-generated citations, under what conditions verification is skipped, and which behavioral or cognitive factors allow…”
Beyond Problem-Solving: The Future of Learning in an AI-Driven World
Burns, L. R. (2025) · Creative Education · doi:10.4236/ce.2025.164031 · Cited by 11
Chatting with Code: Exploring LLMs as Learning Partners in Programming Education
Viberg, O., Wong, J., Feldman-Maggor, Y., Dunder, N., & Epp, C. D. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-98465-5_57 · Cited by 10
The calculator analogy: Epistemic virtues for using LLMs
Voinea, C., Mann, S. P., Savulescu, J., & Earp, B. D. (2025) · Technology in Society · doi:10.1016/j.techsoc.2025.103198 · Cited by 16
ScholarMate: A Mixed-Initiative Tool for Qualitative Knowledge Work and Information Sensemaking
Ye, R., Lee, P. Y. K., Varona, M., Huang, O., & Nobre, C. (2025) · Annual Symposium on Human-Computer Interaction for Work (CHIWORK) · doi:10.1145/3707640.3731913 · Cited by 17
“…risks undermining critical thinking and nuanced interpretation, potentially causing passive acceptance of flawed outcomes [21, 26], thus, ScholarMate is designed as a mixed-initiative system [12], aiming to incorporate AI assistance with human expertise.”
Virtual Collaborative Analysis: Effects of Two AI Summarizers
Kane, A. A., Paletz, S. B. F., Diep, M. M., Hajkowski, A., & Porter, A. (2025) · Small Group Research · doi:10.1177/10464964251361563 · Cited by 7
“…human reliance on AI (Lee & See, 2004) is particularly timely because knowledge workers have self-reported that they often use large language model (LLM)-based AI tools without engaging in the critical thinking needed to identify the inaccuracies that are part of LLM output (Lee et al., 2025).”
AI Literacy and Gender Bias: Comparative Perspectives from the UK and Indonesia
Tunjungbiru, A. D., Pranggono, B., Sari, R. F., Sanchez-Velazquez, E., Purnamasari, P. D., Liliana, D. Y., & Andryani, N. A. C. (2025) · Education Sciences · doi:10.3390/educsci15091143 · Cited by 14
Coscia, A. J., Guo, S., Koh, E., & Endert, A. (2025) · ACM Symposium on User Interface Software and Technology (UIST) · doi:10.1145/3746059.3747746 · Cited by 30
Barus, O., Hidayanto, A., Eitiveni, I., Phusavat, K., & Sudibjo, N. (2025) · Interactive Learning Environments · doi:10.1080/10494820.2025.2596058 · Cited by 9
“Studies even indicate that GenAI has shifted the nature of critical thinking towards merely becoming an information verifier (Lee et al., 2025), and has shown a decline in cognitive abilities for LLM users.”
A Preliminary Set of Principles to Support Learning in the Context of Generative AI
Cooper, M. M. (2025) · Journal of Chemical Education · doi:10.1021/acs.jchemed.5c01413 · Cited by 13
Keeler, J. B., & Murphy, B. A. (2026) · Community Work & Family · doi:10.1080/13668803.2026.2623500 · Cited by 10
Leveling Up or Leveling Down? The Impact of Generative AI on Student Performance in Business Schools
Bergenholtz, C., Vuculescu, O., Günzel-Jensen, F., & Frederiksen, L. F. (2026) · Academy of Management Learning and Education · doi:10.5465/amle.2025.0029 · Cited by 14
Winslow, B. D., Garibay, Ö. Ö., Goyal, T., Koon, S., Margetis, G., Salvendy, G., Shneiderman, B., Tayebi, A., & Vardoulakis, L. (2026) · International Journal of Human-Computer Interaction · doi:10.1080/10447318.2026.2641703 · Cited by 17
Maier, S., Schneider, M., & Feuerriegel, S. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3791185 · Cited by 33
Considering the ethics of large machine learning models in the chemical sciences
Spotte-Smith, E. W. C. (2025) · Machine Learning: Science and Technology · doi:10.1088/2632-2153/adec3c · Cited by 6
“At the same time, foundation model use is associated with reduced critical thinking [76], weakened reasoning, and reduced depth of engagement with information [77], which would seem to suggest that foundation model users are unlikely to take on the added cognitive effort required for bias…”
Large language models and the problem of rhetorical debt
MacArthur, M. (2025) · AI & SOCIETY · doi:10.1007/s00146-025-02403-w · Cited by 8
“Already, research suggests that when knowledge workers place “higher confidence” in AI, they do less critical thinking (Lee et al. 2025).”
Zhu, J., Kempermann, M., Cannanure, V. K., Hartland, A., Navarrete, R. M., Carteny, G., Braun, D., & Weber, I. G. (2025) · ACM Conference on Conversational User Interfaces (CUI) · doi:10.1145/3719160.3736611 · Cited by 8
“Breaking with the trend of AI use cases that reduce users’ critical thinking [33], our deployment showcased an approach to redesign an existing educational tool with LLM to increase cognitive effort while improving user experience.”
Rethinking evidence-informed policy and practice in the age of generative artificial intelligence
Cong‐Lem, N. (2025) · London Review of Education · doi:10.14324/lre.23.1.16 · Cited by 9
“While earlier scholarship (for example, Knox et al., 2022; Lee et al., 2025) has explored CT primarily in relation to learning outcomes or ethical reflection, the EIPP-CT model extends this focus to encompass how educators and policymakers interpret, evaluate and act upon diverse forms of evidence…”
The rise of the research automaton: science as process or product in the era of generative AI?
Sætra, H. S. (2025) · AI & Society · doi:10.1007/s00146-025-02557-7 · Cited by 17
Lee, Y., & Lee, S. (2025) · Sustainability · doi:10.3390/su17177682 · Cited by 22
Mante, D. A., Opoku, O. G., Chineta, O. M., Adamu, A., Parker, D., Xue-fu, X., Hui, X., Setsoafia, N. K., Kyeremeh, P., & Opoku, D. (2025) · Interactive Learning Environments · doi:10.1080/10494820.2025.2556811 · Cited by 7
“…may diminish metacognitive engagement and weaken self-regulation skills, with recent studies suggesting that the frequent use of AI tools in educational settings can lead to cognitive offloading, reducing critical thinking, and impeding reflective practice(Gerlich, 2025; H.-P. H. Lee et al., 2025).”
From limitation to possibility: affordances and tactical evolution in user–GenAI interaction
Zhao, H., Han, Z., Yang, N., & Zeng, Z. (2025) · Journal of Documentation · doi:10.1108/jd-07-2025-0183 · Cited by 8
Reclaiming Agency Through Cyber Humanism: A European Agenda for AI, Education and Culture
Adorni, G., Bellini, E., & Torre, I. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-11108-1_31 · Cited by 7
Held, P., Heubeck, T., & Meckl, R. (2025) · European Journal of Innovation Management · doi:10.1108/ejim-06-2025-0711 · Cited by 9
Addressing Autonomy Risks in Generative Chatbots with the Socratic Method
Lu, W., & Hu, Z. (2025) · Science and Engineering Ethics · doi:10.1007/s11948-025-00567-8 · Cited by 13
“Researchers have found that individuals who outsource cognitive tasks to AI demonstrate significantly lower levels of cognitive engagement and mental effort, indicating a comprehensive takeover of their thought processes by the AI (Zhang & Xu, 2025; Lee et al., 2025).”
Ahmed, A. N., Kataria, S., Saxena, S., & Shankar, A. (2025) · Journal of Consumer Behaviour · doi:10.1002/cb.70086 · Cited by 7
Ferino, S., Hoda, R., Grundy, J., & Treude, C. (2026) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3800580 · Cited by 42
The Digital Double as Counterfeit Spirit: How Externalized Memory Becomes an Autonomous Agent
Pozharov, A. I. (2026) · Философия и культура · doi:10.7256/2454-0757.2026.4.78702 · Cited by 8
Zhu, Z., Yu, J., & Luo, Y. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790697 · Cited by 12
The Soul of the AI: Governance, Ethics, and the Future of Human–AI IntegrationPreprint
Mahajan, P. T. (2025) · doi:10.14293/pr2199.001783.v1 · Cited by 5
Chen, S., Brumby, D., & Cox, A. (2025) · Proceedings of the 4th Annual Symposium on Human-Computer Interaction for Work · doi:10.1145/3729176.3729196 · Cited by 5
Thinking Smarter, not Harder? Google NotebookLM's Misalignment Problem in Education
Albrecht-Crane, C. (2025) · Proceedings of the 43rd ACM International Conference on Design of Communication · doi:10.1145/3711670.3764628 · Cited by 4
Generative AI, cognitive load, and supply chain risk perception: the moderating role of AI literacy
Li, R., Wu, J., Niu, N., & Song, Y. (2025) · International Journal of Logistics Research and Applications · doi:10.1080/13675567.2025.2584305 · Cited by 9
Vrettos, K., Giouroukou, K., Isaac, A., Raissaki, M., & Klontzas, M. E. (2025) · Pediatric Radiology · doi:10.1007/s00247-025-06497-8 · Cited by 7
“Another critical ethical challenge in AI integration is the risk of cognitive biases, where radiologists might either over-rely [101] on or unduly disregard AI suggestions [102].”
Effects of AI explanations on trust and reliance: a study in job shop scheduling
Saßmannshausen, T. M., Burggräf, P., Hassenzahl, M., & Sauer, C. R. (2026) · Ergonomics · doi:10.1080/00140139.2026.2634115 · Cited by 4
Wilson, G., Goodge, T. A., O’Hagan, J., & McGill, M. (2026) · Augmented Humans International Conference (AHs) · doi:10.1145/3795011.3795018 · Cited by 5
“Research has begun to suggest that passively following automated guidance/information from AI leads to poorer understanding of procedural tasks or the content of generated text [49, 99], but giving users prompts to reflect on their use improves knowledge and understanding [94, 99].”
InfluentialDong, A., Li, P., Chen, Y., Gibson, S., Zhao, L., & He, M. (2026) · Sustainability · doi:10.3390/su18115313 · Cited by 9
Co-Designing Collaborative Generative AI Tools for Freelancers
Imteyaz, K., Muller, M. J., Flores-Saviaga, C., & Savage, S. (2026) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3772318.3790280 · Cited by 8
“Recent research on generative AI in group contexts further shows that AI can both support and undermine collaboration, sometimes improving information sharing and coordination, and other times fostering over-reliance that weakens critical discussion [79, 112, 165, 216].”
InfluentialA scoping review on the mental health harms of LLM-based chatbots
Diel, A., Torous, J., Cuijpers, P., Kleesiek, J., Nensa, F., Weber, N., Faust, F., Lalgi, T. J., Mellis, F. S., Teufel, M., et al. (2026) · npj Digital Medicine · doi:10.1038/s41746-026-03054-x · Cited by 4
“One study found a decrease in cognitive functioning related to the task at hand, alongside gainsincriticalthinkingrelatedtosupervisorytasks(e.g.,veri fi cationand stewardship of LLM-CB outputs), indicating a cognitive adaptation integrating LLM-CBs into cognitive problem solving 93 .”
How AI Impacts Skill Formation
Shen, J. H., & Tamkin, A. (2026) · arXiv · doi:10.48550/arxiv.2601.20245 · Cited by 120
“Questions that include both code generation and explanation exhibiting these interaction patterns showed less independent thinking and more cognitive offloading [Lee et al., 2025].”
Papadakis, S. (2025) · Journal of Baltic Science Education · doi:10.33225/jbse/25.24.592 · Cited by 3
AI-Informed Pedagogy for a Post-Truth Era
Nally, D. (2025) · Digital Society · doi:10.1007/s44206-025-00230-6 · Cited by 5
“An even more recent case study (Lee et al., 2025 ) echoes some of these findings, although its focus was on degree of perceived effort exerted in critical thinking exercises, in the collabora - tive use of AIs by 319 knowledge workers.”
Rethinking what is valuable in mathematics and statistics education
Horrocks, M., & Shearman, D. (2025) · International Journal of Mathematical Education in Science and Technology · doi:10.1080/0020739x.2025.2556864 · Cited by 7
“…systems is creating much discussion about the ways in which ‘gener-ative AI will transform knowledge work’ (AbuMusab, 2024; Alavi & Westerman, 2023; Lee et al., 2025; Storey et al., 2025) and claims about the skills that will be rendered obsolete and those that will still be important (AbuMusab,…”
Exposure to content written by large language models can reduce stigma around opioid use disorder
Mittal, S., Shah, D., Do, S. W., ElSherief, M., Mitra, T., & Choudhury, M. D. (2025) · npj Artificial Intelligence · doi:10.1038/s44387-025-00049-z · Cited by 7
“We recruited participants through Proli fi c, a crowd-sourcing platform extensively used in computational social science 67,68 .”
Mohale, N. E., Sevnarayan, K., Maphoto, K. B., & Suliman, Z. (2025) · Online Journal of Communication and Media Technologies · doi:10.30935/ojcmt/17484 · Cited by 5
Hugerth, M. W., & Hugerth, L. W. (2025) · Computers and Education Artificial Intelligence · doi:10.1016/j.caeai.2025.100531 · Cited by 10
Balcerzak, A. P., Zinecker, M., & Mičánek, J. (2025) · Human Technology · doi:10.14254/1795-6889.2025.21-3.7 · Cited by 13
Metadata
- Authors
- Lee, H.-P. (., Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N.
- Affiliations
- Carnegie Mellon University (US)
- Microsoft Research (United Kingdom) (GB)
- ORCID
- Venue
- Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems
- Publisher
- ACM
- Published
- 2025-04-25
- Pages
- 1-22
- OpenAlex
- W4409885395
- Cited by
- 490
- References
- 118
- Citations by year
- 2026: 328 · 2025: 166
- FWCI
- 573.88
- Type
- conference paper
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- CHI Conference on Human Factors in Computing Systems (publishedVersion, cc-by)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Intelligent Tutoring Systems and Adaptive Learning
- Social Sciences › Education › Education and Critical Thinking Development
- Social Sciences › Business, Management and Accounting › Marketing › AI in Service Interactions
- Keywords
- generative AI, critical thinking, knowledge workers, self-confidence, cognitive effort, fact-checking, response integration, AI tool design
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-09-28, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-09-28 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.

