Pith. sign in

REVIEW 3 major objections 7 minor 1 cited by

Do AI tutors empower or enslave learners? Toward a critical use of AI in education

T0 review · 3 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This position paper claims that unchecked AI use in education risks cognitive atrophy, loss of agency, emotional harm, and privacy violations, and that intentional, learning-science-based design makes AI empowering rather than enslaving.

desk verdict A coherent, well-written position paper that consolidates known risks of AI in education but adds no new evidence; its main soft spot is reading a correlational study as proof of long-term cognitive decline, yet the paper hedges enough to warrant engagement. read the letter →

arxiv 2507.06878 v1 pith:3W2ZP3YS submitted 2025-07-09 cs.CY cs.HC

Lucile Favero , Juan-Antonio Pérez-Ortiz , Tanja Käser , Nuria Oliver This is my paper
classification cs.CYcs.HC
keywords cognitiveoffloadingcriticalthinkingAIineducationlearneragencydesirabledifficultiesliteracyacademicintegritystudentwell-being
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This position paper argues that the same AI tools that can tutor and personalize learning can, when used without guardrails, weaken the very capacities education exists to build: independent critical thinking, memory, creativity, deep reasoning, and the motivation to exert cognitive effort. Drawing on cognitive science and pedagogy, it claims that generic chatbots encourage cognitive offloading of core thinking tasks, fostering dependency, passivity, conformity, anxiety and 'AI guilt', while also raising privacy and academic-integrity concerns. The paper does not oppose AI in education; it argues that the outcome depends on design and use. AI becomes empowering when it scaffolds reasoning and preserves desirable difficulties, and enslaving when it replaces the learner's own cognitive engagement. A sympathetic reader would care because this reframes the AI-in-education debate from tool adoption to a design and policy choice.

What carries the argument

The mechanism carrying the argument is the distinction between cognitive offloading that frees mental resources for deeper thinking and offloading that replaces core analytical, reasoning, and synthesis tasks. The paper's risk model organises harms along four axes: cognition (atrophy, dependency, conformity), agency (loss of autonomy, over-trust), emotion (self-efficacy, AI guilt, technostress), and ethics (privacy, surveillance, data exploitation, academic integrity). Against this it sets pedagogical levers: scaffolding, active learning, metacognitive training, desirable difficulties, and AI literacy. These concepts do the work of both diagnosing why generic chatbots undercut learning and prescribing why well-designed AI tutors succeed.

What would settle it

A semester-long randomised experiment that assigns students to three conditions — a generic chatbot, a theory-designed AI tutor with scaffolding and desirable difficulties, and no AI — and measures Halpern Critical Thinking Assessment scores before and after would settle the claim: if the generic-chatbot group shows no decline relative to the no-AI control after controlling for baseline scores, the cognitive-atrophy claim fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that over-reliance on AI, especially chatbots, may produce cognitive atrophy: reduced independent and critical thinking, memory, creativity, deep reasoning, and motivation to exert cognitive effort, along with loss of agency, emotional harm (lowered self-efficacy, AI guilt, cognitive dissonance, technostress), and ethical and privacy violations. It distinguishes between AI that acts as a cognitive shortcut, delivering fast pre-digested answers that outsource thinking, and AI that acts as a scaffold, supporting goal-setting, metacognition, and effortful processing. Evidence cited includes a user study linking higher AI reliance to lower scores on the Halpern Critical Thinking Assessment mediated by cognitive offloading, and a study showing theory-designed AI tutors outperforming active learning in both learning and time. The paper argues that aligning AI with active learning, constructivism, scaffolding, and desirable difficulties is what makes the difference, and it advocates for human-centered AI that empowers learners.

Load-bearing premise

The whole argument leans on the premise that the observed correlation between higher AI reliance and lower critical-thinking scores reflects a causal, long-term cognitive decline, rather than pre-existing differences among students or short-lived task effects.

Editorial extensions

If this is right

  • If the paper is right, educational institutions should treat AI literacy as a core competency, not an optional add-on, since only 25% of U.S. colleges currently offer formal AI training.
  • AI tools should be designed or prompted to scaffold reasoning: asking questions, delaying answers, and comparing AI outputs with human logic rather than giving direct solutions.
  • Assessments need redesign to reward critical engagement and make academic integrity clear, because generic chatbots resist traditional plagiarism detection.
  • Institutions need privacy and consent frameworks that protect students' right to make mistakes without surveillance, and that comply with regulations like the GDPR.
  • Curricula should teach how AI systems work, demystifying them, to reduce over-trust and support students' confidence in their own abilities.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's framework implies a testable design principle: in a randomised semester-long trial, an AI tutor built on desirable difficulties and Socratic prompting should produce larger gains in critical-thinking assessments than a generic chatbot, with lower AI guilt.
  • A longitudinal study that tracks the same students before and after AI adoption, controlling for baseline critical-thinking scores and study habits, would be needed to confirm that cognitive atrophy is a causal effect of offloading rather than a selection effect.
  • The paper's 'AI guilt' analysis suggests that clear acceptable-use norms might reduce cognitive dissonance and emotional harm, a prediction that could be tested by comparing institutions with explicit AI policies against those without.
  • If over-trust stems from anthropomorphism, then interface changes that make AI outputs less conversational and more machine-like could reduce uncritical acceptance, a hypothesis the paper gestures at but does not test.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 7 minor

Summary. The paper is a position paper that argues that unchecked use of AI, particularly chatbots, in education may lead to cognitive atrophy, loss of learner agency, emotional harm, and ethical and privacy violations. It synthesizes literature from cognitive science and pedagogy, distinguishes generic chatbots from pedagogically designed AI tutors, and proposes three principles—empowering reasoning, fostering emotional well-being, and institutional support—for a critical and human-centered use of AI in education. The central claim is hedged with hedges such as 'may' and 'can,' but the argumentative structure, especially the contrast between 'enslaving' and 'empowering' AI use, requires a causal reading of the cited evidence.

Significance. If its central claim is treated as a clearly labeled hypothesis rather than an established finding, the paper offers a timely and useful synthesis. It organizes a wide range of concerns into a coherent framework (cognition, agency, emotion, ethics), brings in relevant cognitive science principles such as cognitive offloading and desirable difficulties, and provides actionable recommendations, including scaffolding, metacognitive training, and AI literacy. The distinction between generic chatbots and well-designed AI tutors is conceptually valuable and grounded in a concrete example. The paper does not present original empirical data, but that in itself is not a weakness for a position paper. The main risk is that the risk model's empirical foundation is weaker than the prose suggests: several load-bearing claims rest on correlational, cross-sectional, or self-report evidence that is presented with causal language. Addressing this gap would make the paper's cautionary message more defensible and more useful to educators and policymakers.

major comments (3)
  1. [Section 2.1] The central claim that AI use 'can erode' independent reasoning is based on Gerlich [9], which is a concurrent correlational study. The sentence 'Over-reliance on AI may lead to reduced independent and critical thinking' is appropriately hedged, but the subsequent sentence 'automation of analytical tasks can erode the learners' ability' and the broader risk model depend on a causal relationship between AI reliance and long-term cognitive decline. Because the HCTA score and self-reported AI reliance are measured at the same time, the correlation could reflect selection (learners with weaker critical thinking using AI more heavily) or short-term task effects rather than atrophy. Please reframe this as an association and add an explicit call for longitudinal and interventional studies that manipulate AI reliance and measure critical-thinking outcomes over time.
  2. [Section 2.2] The 'empowering' side of the central distinction is supported by the claim that 'well-designed AI tutors can outperform even active learning techniques,' which cites only a non-peer-reviewed preprint (Kestin et al. [24], published on Research Square). This is a load-bearing point because it justifies the contrast between generic chatbots and theory-driven AI tutors. Given the paper's evidence-based framing, this claim should be explicitly labeled as a preprint with its peer-review status noted, and ideally corroborated by additional published studies. Alternatively, the authors should temper the claim to reflect that the evidence is preliminary.
  3. [Section 4.1] The causal cycle described in Section 4.1—'the more students use AI to avoid academic challenges, the less confident they become in their own abilities'—is presented as established fact, but the cited studies [5, 12, 16, 37] are largely cross-sectional, qualitative, or self-report based. This same pattern of overreach appears elsewhere, for instance in Section 3.1 ('excessive use of AI can deteriorate emotional intelligence and human connection' [32]). The paper would benefit from an explicit limitations paragraph, perhaps in Section 7 or 8, acknowledging that many of the cited studies cannot distinguish causal effects from pre-existing differences or reverse causality, and that the risk model should be read as a set of hypotheses to be tested.
minor comments (7)
  1. [Introduction] There are two typos: 'human-centic' should be 'human-centric,' and the phrase 'inadequately protected make [13]' in the last paragraph should be rewritten (likely 'inadequately protected, as noted in [13]').
  2. [Section 5.1 / Introduction] The phrase 'make [13]' in the Introduction is grammatically incomplete; please revise the sentence for clarity.
  3. [Section 6.1] The reference [42] to the Digital Education Council is given only as a URL; please supply the full report title, publication date, and authoring organization so that readers can verify the survey details.
  4. [References] Reference [9] is incomplete: 'MDPI (2025)' does not identify the journal name, volume, article number, or DOI. Please complete the bibliographic information.
  5. [Section 3.1] The sentence 'Even when AI is used in conjunction with metacognitive scaffolds, students often fail to regain full autonomy' is a strong empirical claim with no supporting citation; please add a source or soften the wording.
  6. [Conclusion] The final sentence contains a grammatical error: 'tools that we entrust to that help us with that task' should be 'tools that we entrust to help us with that task.'
  7. [Figure 1] Figure 1 is central to the paper's taxonomy but is not described in the body text; please add a brief walk-through of the boxes and arrows so that readers can interpret the figure without guessing.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; the argument is a synthesis of external literature, with one minor self-citation that is not load-bearing.

full rationale

The paper is a position paper that assembles external evidence from cognitive science, education, and empirical studies; it does not fit parameters and then rename them as predictions, and it contains no derivation chain whose conclusion is equivalent to its premises by construction. The central risk claim (Section 2.1) that over-reliance on AI 'may lead to reduced independent and critical thinking' is supported by external sources, notably Gerlich's correlational HCTA study [9]; the paper's move from correlation to the word 'erode' is a causal-inference concern about evidence strength, not a circularity, because the cited study is independent of the paper's own claims. The only self-citation, [43], appears in Section 7 in the phrase 'AI tools should be used to support scaffolding, prompting deeper reflection, reasoning, and understanding [2, 36, 43].' This is a minor, non-load-bearing reference supporting an optional design recommendation; the paper's warning about cognitive atrophy, loss of agency, emotional harm, and privacy risks does not depend on the authors' prior work. No uniqueness theorem or ansatz is imported from the authors' own publications, and no known empirical result is merely renamed. The paper is self-contained against external benchmarks in the sense that its claims are checkable against the cited literature. Accordingly, the circularity burden is low; the score of 2 reflects only the presence of the minor self-citation.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new entities, forces, or parameters. Its argument is built on established educational psychology principles and empirical studies from the literature. The sections above-ground each assumption serve as the theoretical foundation for the claims.

assumptions (4)
  • domain assumption Active learning and constructivist methods are more effective than passive reception.
    Invoked in Section 2.2, building on Freeman et al. [20] and constructivist theory [21], to argue that AI should promote active engagement.
  • domain assumption Desirable difficulties (spacing, interleaving, retrieval, delayed feedback) improve long-term retention.
    Used in Section 2.3 to argue that AI tools should preserve mental effort and avoid overly fluent answers.
  • domain assumption Cognitive offloading can either support or undermine learning depending on what is offloaded.
    Section 2.1 relies on Risko and Gilbert [17] to distinguish beneficial offloading from offloading core thinking tasks.
  • domain assumption Student self-reported concerns and attitudes are reliable indicators of learning outcomes and well-being.
    Section 6 uses surveys and qualitative comments from [9, 42] as evidence for emotional and motivational harms.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Do AI tutors empower or enslave learners? Toward a critical use of AI in education." pith.science (2026). https://pith.science/paper/3W2ZP3YS

@misc{pith2026250706878,
  author       = {Pith},
  title        = {Pith review of: Do AI tutors empower or enslave learners? Toward a critical use of AI in education},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3W2ZP3YS}},
  note         = {Machine review of arXiv:2507.06878}
}
read the original abstract

The increasing integration of AI tools in education presents both opportunities and challenges, particularly regarding the development of the students' critical thinking skills. This position paper argues that while AI can support learning, its unchecked use may lead to cognitive atrophy, loss of agency, emotional risks, and ethical concerns, ultimately undermining the core goals of education. Drawing on cognitive science and pedagogy, the paper explores how over-reliance on AI can disrupt meaningful learning, foster dependency and conformity, undermine the students' self-efficacy, academic integrity, and well-being, and raise concerns about questionable privacy practices. It also highlights the importance of considering the students' perspectives and proposes actionable strategies to ensure that AI serves as a meaningful support rather than a cognitive shortcut. The paper advocates for an intentional, transparent, and critically informed use of AI that empowers rather than diminishes the learner.

Figures

Figures reproduced from arXiv: 2507.06878 by the authors.

Figure 1
Figure 1. Ensuring a Human-Centered, responsible, and critical use of AI in education: challenges and interventions syndrome, cognitive dissonance, or digital isolation [12]. Data privacy is another concern as student data could po￾tentially be misused or inadequately protected make [13]. In this paper, we reflect on and discuss these ambivalent implications of AI integration in education, with a focus on the impact of chatbo… view at source ↗

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. LLM Harms: A Taxonomy and Discussion

    cs.CY 2025-12 unverdicted novelty 3.0 of 10

    This paper proposes a taxonomy of LLM harms in five categories and suggests mitigation strategies plus a dynamic auditing system for responsible development.

Reference graph

Works this paper leans on

47 extracted references · 39 canonical work pages · cited by 1 Pith paper

  1. [9]

    Gerlich, AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking, MDPI (2025)

    M. Gerlich, AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking, MDPI (2025)

    work page 2025

  2. [24]

    Kestin*, K

    G. Kestin*, K. Miller*, A. Klales, T. Milbourne, G. Ponti, AI Tutoring Outperforms Active Learning,

    work page

  3. [32]

    Klimova, M

    B. Klimova, M. Pikhart, Exploring the effects of artifi- cial intelligence on student and academic well-being in higher education: a mini-review, Frontiers in Psy- chology 16 (2025) 1498132

    work page 2025

  4. [1]

    Accessed: 2025-05-25

    UNESCO, Artificial intelligence in education, https://www.unesco.org/en/digital-education/ artificial-intelligence, 2025. Accessed: 2025-05-25

    work page 2025

  5. [2]

    Létourneau, M

    A. Létourneau, M. Deslandes Martineau, P. Char- land, J. A. Karran, J. Boasen, P. M. Léger, A systematic review of AI-driven intelligent tutor- ing systems (ITS) in K-12 education, npj Sci- ence of Learning 10 (2025) 1–13. URL: https:// www.nature.com/articles/s41539-025-00320-7. doi: 10. 1038/s41539-025-00320-7

    work page 2025

  6. [3]

    Tafazoli, Exploring the potential of genera- tive AI in democratizing English language educa- tion, Computers and Education: Artificial Intelli- gence 7 (2024) 100275

    D. Tafazoli, Exploring the potential of genera- tive AI in democratizing English language educa- tion, Computers and Education: Artificial Intelli- gence 7 (2024) 100275. URL: https://linkinghub.elsevier. com/retrieve/pii/S2666920X2400078X. doi:10.1016/ j.caeai.2024.100275

    work page arXiv 2024

  7. [4]

    A. M. Al-Zahrani, T. M. Alasmari, Exploring the impact of artificial intelligence on higher educa- tion: The dynamics of ethical, social, and educa- tional implications, Humanities and Social Sci- ences Communications 11 (2024) 1–12. URL: https:// www.nature.com/articles/s41599-024-03432-4. doi: 10. 1057/s41599-024-03432-4

    work page 2024

  8. [5]

    M. Lan, X. Zhou, A qualitative system- atic review on AI empowered self-regulated learning in higher education, npj Science of Learning 10 (2025) 1–16. URL: https: //www.nature.com/articles/s41539-025-00319-0. doi:10.1038/s41539-025-00319-0

    doi:10.1038/s41539-025-00319-0 2025

Show all 47 references
  1. [6]

    K. I. Vorobyeva, S. Belous, N. V. Savchenko, L. M. Smirnova, S. A. Nikitina, S. P. Zhdanov, Personalized learning through AI: Pedagogical approaches and crit- ical insights, Contemporary Educational Technology 17 (2025) ep574. doi:10.30935/cedtech/16108

    2025 doi

  2. [7]

    Bank, Ai: The software can instantly grade,

    W. Bank, Ai: The software can instantly grade,

  3. [8]

    D. e. a. Ravšelj, Higher education students’ perceptions of ChatGPT: A global study of early reactions, PLOS ONE 20 (2025) e0315011. URL: https://journals.plos. org/plosone/article?id=10.1371/journal.pone.0315011. doi:10.1371/journal.pone.0315011

    2025 doi

  4. [10]

    Darvishi, H

    A. Darvishi, H. Khosravi, S. Sadiq, D. Gašević, G. Siemens, Impact of ai assistance on student agency, Computers & Education 210 (2024) 104967

    2024

  5. [11]

    J. Roe, M. Perkins, Generative ai and agency in educa- tion: A critical scoping review and thematic analysis, arXiv preprint arXiv:2411.00631 (2024)

    2024 arXiv

  6. [12]

    Rodríguez-Ruiz, I

    J. Rodríguez-Ruiz, I. Marín-López, R. Espejo- Siles, Is artificial intelligence use related to self-control, self-esteem and self-efficacy among university students?, Education and Information Technologies 30 (2025) 2507–2524. URL: https://doi.org/10.1007/s10639-024-12906-6. d...

    2025 doi

  7. [13]

    M. B. Kwapisz, A. Kohli, P. Rajivan, Privacy Con- cerns of Student Data Shared with Instructors in an Online Learning Management System, in: Proceed- ings of the 2024 CHI Conference on Human Factors in Computing Systems, CHI ’24, Association for Com- puting Machinery, 2024, pp...

    2024

  8. [14]

    C. Zhai, S. Wibowo, L. D. Li, The effects of over- reliance on ai dialogue systems on students’ cognitive abilities: a systematic review, Smart Learning Envi- ronments 11 (2024) 28

    2024

  9. [15]

    R. T. Williams, The ethical implications of using gen- erative chatbots in higher education, in: Frontiers in Education, volume 8, Frontiers Media SA, 2024, p. 1331607

    2024

  10. [16]

    Zhang, X

    S. Zhang, X. Zhao, T. Zhou, J. H. Kim, Do you have AI dependency? The roles of academic self-efficacy, aca- demic stress, and performance expectations on prob- lematic AI usage behavior, International Journal of Educational Technology in Higher Education 21 (2024)

    2024

  11. [17]

    E. F. Risko, S. J. Gilbert, Cognitive Offloading, Trends in Cognitive Sciences 20 (2016) 676–688. URL: https://www.cell.com/trends/cognitive-sciences/ abstract/S1364-6613(16)30098-5. doi: 10.1016/j. tics.2016.07.002. arXiv:27542527

    2016 doi

  12. [18]

    D. F. Halpern, The halpern critical thinking assess- ment: Manual, Modling, Austria: Schuhfried GmbH (2010)

    2010

  13. [19]

    Mezirow, Transformative Learning: Theory to Prac- tice, New Directions for Adult and Continuing Educa- tion 1997 (1997) 5–12

    J. Mezirow, Transformative Learning: Theory to Prac- tice, New Directions for Adult and Continuing Educa- tion 1997 (1997) 5–12. URL: https://onlinelibrary.wiley. com/doi/10.1002/ace.7401. doi:10.1002/ace.7401

    1997 doi

  14. [20]

    Freeman, S

    S. Freeman, S. L. Eddy, M. McDonough, M. K. Smith, N. Okoroafor, H. Jordt, M. P. Wenderoth, Active learn- ing increases student performance in science, engi- neering, and mathematics, Proceedings of the national academy of sciences 111 (2014) 8410–8415

    2014

  15. [21]

    C. T. Fosnot, Constructivism: Theory, perspectives, and practice, Teachers College Press, 2013

    2013

  16. [22]

    R. R. Van Der Stuyf, Scaffolding as a teaching strategy, Adolescent learning and development 52 (2002) 5–18

    2002

  17. [23]

    URL: https://cdn.openai.com/global-affairs/ openai-edu-ai-ready-workforce.pdf, online; January 2025

    OpenAI, Building an AI-Ready Workforce: A Look at College Student ChatGPT Adoption in the US, Technical Report, OpenAI, 2025. URL: https://cdn.openai.com/global-affairs/ openai-edu-ai-ready-workforce.pdf, online; January 2025

    2025

  18. [25]

    A. M. Al-Zahrani, Unveiling the shad- ows: Beyond the hype of AI in education, Heliyon 10 (2024). URL: https://www.cell. com/heliyon/abstract/S2405-8440(24)06727-6. doi:10.1016/j.heliyon.2024.e30696

    2024 doi

  19. [26]

    E. L. Bjork, R. A. Bjork, et al., Making things hard on yourself, but in a good way: Creating desirable diffi- culties to enhance learning, Psychology and the real world: Essays illustrating fundamental contributions to society 2 (2011)

    2011

  20. [27]

    doi: 10.21203/rs.3.rs-4243877/ v1

    URL: https://www.researchsquare.com/article/ rs-4243877/v1. doi: 10.21203/rs.3.rs-4243877/ v1

    doi

  21. [28]

    Y. Mera, G. Rodríguez, E. Marin-Garcia, Unraveling the benefits of experiencing errors during learning: Definition, modulating factors, and explanatory theo- ries, Psychonomic Bulletin & Review 29 (2022) 753–

    2022

  22. [29]

    Deslauriers, L

    L. Deslauriers, L. S. McCarty, K. Miller, K. Callaghan, G. Kestin, Measuring actual learning versus feel- ing of learning in response to being actively en- gaged in the classroom, Proceedings of the National Academy of Sciences 116 (2019) 19251– 19257. URL: https://www.pnas.or...

    2019 doi

  23. [30]

    A. B. de Bruin, F. Biwer, L. Hui, E. Onan, L. David, W. Wiradhany, Worth the effort: The start and stick to desirable difficulties (s2d2) framework, Educational Psychology Review 35 (2023) 41

    2023

  24. [31]

    M. H. Howlader, T. Mortuja Mahamud, Z. Say- eeda, C. Sanjoy Kumar, J. Guo, Factors influ- encing the acceptance and usage of ChatGPT as an emerging learning tool among higher educa- tion students in Bangladesh: A structural equation modeling, Cogent Education 12 (2025) 2504224...

    2025

  25. [33]

    Bećirović, E

    S. Bećirović, E. Polz, I. Tinkel, Exploring students’ ai literacy and its effects on their ai output quality, self- efficacy, and academic performance, Smart Learning Environments 12 (2025) 29

    2025

  26. [34]

    doi:10.1186/s41239-024-00467-0

    URL: https://doi.org/10.1186/s41239-024-00467-0. doi:10.1186/s41239-024-00467-0

    doi

  27. [35]

    Biwer, A

    F. Biwer, A. B. de Bruin, S. Schreurs, M. G. oude Eg- brink, Future steps in teaching desirably difficult learn- ing strategies: Reflections from the study smart pro- gram, Journal of Applied Research in Memory and Cognition 9 (2020) 439–446

    2020

  28. [36]

    Bastani, O

    H. Bastani, O. Bastani, A. Sungu, H. Ge, O. Kabakci, R. Mariman, Generative AI Can Harm Learning, 2024. URL: https://www.ssrn.com/abstract=4895486. doi:10. 2139/ssrn.4895486

    2024

  29. [37]

    AI Guilt

    C. K. Y. Chan, Exploring the Factors of "AI Guilt" Among Students – Are You Guilty of Using AI in Your Homework?, 2024. URL: http://arxiv.org/ abs/2407.10777. doi:10.48550/arXiv.2407.10777. arXiv:2407.10777

    2024 doi arXiv

  30. [38]

    Gerlich, The Power of Virtual Influencers: Impact on Consumer Behaviour and Attitudes in the Age of AI, Administrative Sciences 13 (2023) 178

    M. Gerlich, The Power of Virtual Influencers: Impact on Consumer Behaviour and Attitudes in the Age of AI, Administrative Sciences 13 (2023) 178. URL: https: //www.mdpi.com/2076-3387/13/8/178. doi:10.3390/ admsci13080178

    2023

  31. [39]

    Schaaff, M.-A

    K. Schaaff, M.-A. Heidelmann, Impacts of anthropo- morphizing large language models in learning envi- ronments, arXiv preprint arXiv:2408.03945 (2024)

    2024 arXiv

  32. [40]

    Mason, E

    A. Mason, E. Yerushalmi, E. Cohen, C. Singh, Learning from mistakes: The effect of students’ written self- diagnoses on subsequent problem solving, The Physics Teacher 54 (2016) 87–90

    2016

  33. [41]

    A. C. Niloy, R. Hafiz, B. M. Hossain, F. Gulmeher, N. Sultana, K. F. Islam, F. Bushra, S. Islam, S. I. Hoque, M. Rahman, S. Kabir, AI chatbots: A disguised enemy for academic integrity?, International Jour- nal of Educational Research Open 7 (2024) 100396. URL: http://www.scie...

    2024

  34. [42]

    A. J. Elliot, P. G. Devine, On the motivational nature of cognitive dissonance: Dissonance as psychological dis- comfort., Journal of personality and social psychology 67 (1994) 382

    1994

  35. [43]

    Shores, Surveillance Education: Tracks the Rise of Spying Technology in Schools, MIT Press, 2024

    T. Shores, Surveillance Education: Tracks the Rise of Spying Technology in Schools, MIT Press, 2024

    2024

  36. [46]

    Ac- cessed: 2025-05-27

    Digital Education Council, Digital education council, https://www.digitaleducationcouncil.com, 2025. Ac- cessed: 2025-05-27

    2025

  37. [47]

    Favero, J

    L. Favero, J. A. Pérez-Ortiz, T. Käser, N. Oliver, En- hancing critical thinking in education by means of a socratic chatbot, arXiv preprint arXiv:2409.05511 (2024)

    2024 arXiv

  38. [765]

    doi:10.3758/s13423-021-02022-8

    URL: https://doi.org/10.3758/s13423-021-02022-8. doi:10.3758/s13423-021-02022-8

    doi

  39. [2024]

    org/curated/en/099734306182493324/pdf/ IDU152823b13109c514ebd19c241a289470b6902.pdf, accessed: 2025-05-25

    URL: https://documents1.worldbank. org/curated/en/099734306182493324/pdf/ IDU152823b13109c514ebd19c241a289470b6902.pdf, accessed: 2025-05-25

    2025

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.