arxiv: 2605.09624 · v1 · submitted 2026-05-10 · ⚛️ physics.ed-ph · cs.CY· physics.app-ph

Recognition: 2 theorem links

· Lean Theorem

Preparing Students for AI-Powered Materials Discovery: A Workflow-Aligned Framework for AI Literacy, Equity, and Scientific Judgment

Ben Sayler, Dongming Mei, Katherine Moore

Authors on Pith no claims yet

Pith reviewed 2026-05-12 03:23 UTC · model grok-4.3

classification ⚛️ physics.ed-ph cs.CYphysics.app-ph

keywords AI literacymaterials discoveryscientific judgmentworkflow-aligned educationmaterials informatics competenciesequity in learning outcomescognitive off-loadingcurriculum framework

0 comments

The pith

AI education for materials discovery must move beyond tool access to a workflow-aligned literacy model that builds scientific judgment.

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

The paper contends that AI is transforming materials science through property prediction and hypothesis generation, but effective training now depends on whether students can apply AI with domain-specific judgment rather than treating it as a black-box tool. It establishes that current approaches emphasizing surface interaction with generative systems fall short and must be replaced by a framework that embeds AI literacy within the full research workflow, linking it explicitly to competencies such as data provenance, featurization, validation, uncertainty handling, and experimental feedback. If this shift occurs, programs could produce students who maintain independent scientific reasoning while using AI, achieve more uniform learning outcomes across diverse groups, and avoid pitfalls like over-reliance that undermine research quality.

Core claim

The paper's central claim is that AI literacy for materials discovery must be workflow-aligned, meaning it directly connects literacy development to materials-informatics competencies including data provenance, domain-specific featurization, model validation, uncertainty quantification, physics-informed reasoning, reproducibility, and experimental feedback, while also tracking outcome-oriented equity metrics such as comparable gains, confidence calibration, persistence, and research readiness across subgroups and mitigating risks including cognitive off-loading and cognitive surrender through a dual-track curriculum model suitable for courses, bootcamps, workshops, and program reform.

What carries the argument

The workflow-aligned model of AI literacy, which integrates AI capabilities into the complete sequence of materials research steps from data handling through validation and iteration rather than isolating tool use.

If this is right

Students gain the ability to apply AI while preserving independent scientific reasoning across the discovery process.
Educational programs achieve comparable learning gains, transfer, and research readiness for all student subgroups.
Risks of cognitive off-loading and surrender decrease as AI use is tied to validation and feedback loops.
Dual-track curriculum structures become implementable in courses, bootcamps, and full programs with associated assessment plans.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The framework's emphasis on physics-informed reasoning could extend to training in adjacent fields where AI assists experiment design, such as chemistry or biology.
Institutions adopting the model may need to revise assessment rubrics to explicitly score confidence calibration alongside task performance.
Early exposure to workflow-aligned literacy in undergraduate programs could reduce later remediation needs when students enter AI-heavy research labs.

Load-bearing premise

That implementing this workflow-aligned model connected to materials-informatics competencies will produce better scientific judgment, equitable outcomes across subgroups, and fewer risks such as cognitive off-loading, even though the paper supplies no empirical tests of these effects.

What would settle it

A controlled comparison of student cohorts using the proposed curriculum versus standard tool-access training, measuring scientific judgment via tasks requiring AI-assisted hypothesis evaluation and tracking subgroup differences in learning gains and confidence calibration, that finds no measurable improvements.

Figures

Figures reproduced from arXiv: 2605.09624 by Ben Sayler, Dongming Mei, Katherine Moore.

**Figure 2.** Figure 2: Dual-track approach to AI education. Foundational AI courses and disciplinary [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

**Figure 3.** Figure 3: Benefits and risks of AI-supported education. Educational value is maximized [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

**Figure 4.** Figure 4: Integrated strategy set for AI education in materials informatics. Governance [PITH_FULL_IMAGE:figures/full_fig_p014_4.png] view at source ↗

read the original abstract

Artificial intelligence (AI) is reshaping education, scientific training, and materials discovery. In materials science, AI models increasingly support property prediction, experiment prioritization, and hypothesis generation; however, the limiting factor is no longer only algorithmic capability but also whether students and educators can use AI with domain-specific scientific judgment. This workshop-informed white paper and curriculum-oriented position article argues that AI education for AI-powered materials discovery must move beyond tool access and surface-level interaction with generative AI systems toward a workflow-aligned model of AI literacy. We connect AI literacy to materials-informatics competencies: data provenance, domain-specific featurization, model validation, uncertainty quantification, physics informed reasoning, reproducibility, and experimental feedback. We also emphasize outcome-oriented equity: institutions should evaluate not only access, participation, and engagement, but also whether AI-enabled instruction produces comparable learning gains, transfer of learning, confidence calibration, defined as the alignment with students confidence and the quality or correctness of their work, persistence, and research readiness across student subgroups. The paper synthesizes relevant evidence, identifies risks for learners such as cognitive off-loading and cognitive surrender, and provides a dual-track curriculum model and implementation recommendations such as curriculum guides and an assessment plan for courses, bootcamps, workshops, and program-level reform. The central goal is to prepare students to become better scientists, not merely more efficient users of AI tools.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This is a position paper proposing a workflow-tied AI literacy curriculum for materials science students, but it offers no data or tests to show the approach works better than simpler alternatives.

read the letter

The core takeaway is that AI training in materials discovery needs to go beyond tool access and link directly to domain tasks like data provenance, uncertainty handling, and physics-informed checks. The authors lay out a dual-track model plus equity checks and risk warnings such as cognitive off-loading. That framing is reasonable and pulls together scattered ideas from AI literacy work and materials informatics into one place with implementation notes like assessment outlines for courses and workshops. It gives educators a concrete structure to consider rather than vague calls for more AI exposure. The synthesis is clear and the equity focus on comparable gains across subgroups is a useful reminder. The main limitation is that none of the claimed benefits—better scientific judgment, reduced risks, or equitable outcomes—are backed by any original evidence in the paper. There are no pilots, pre/post results, or controlled comparisons; the argument stays at the level of reasoned proposal drawing on external citations. That makes the prescriptive tone feel ahead of the supporting material. The piece is aimed at curriculum designers and program leads in materials science or adjacent STEM fields who are already thinking about AI integration. A reader looking for discussion starters or a checklist of competencies will get value, while someone seeking validated methods or new measurements will not. It is coherent on its own terms and engages the literature without internal contradictions, so it deserves a serious referee at a venue open to position or white papers. Reviewers can push for added examples or references to existing programs, but the topic is timely enough to warrant that step rather than a desk rejection.

Referee Report

2 major / 2 minor

Summary. This workshop-informed white paper and position article argues that AI education for materials discovery must shift from tool access and surface-level generative AI use to a workflow-aligned model of AI literacy. It connects this literacy to materials-informatics competencies (data provenance, domain-specific featurization, model validation, uncertainty quantification, physics-informed reasoning, reproducibility, and experimental feedback). The paper stresses outcome-oriented equity (comparable learning gains, transfer, confidence calibration, persistence, and research readiness across subgroups), synthesizes external evidence, identifies risks such as cognitive off-loading and cognitive surrender, and proposes a dual-track curriculum model plus implementation recommendations (curriculum guides, assessment plans) for courses, bootcamps, workshops, and program reform. The goal is to prepare students as better scientists rather than efficient AI tool users.

Significance. If the proposed workflow-aligned framework is adopted and subsequently validated, it would hold substantial significance for materials science and physics education by promoting domain-specific scientific judgment alongside AI tools and addressing equity in outcomes. The synthesis of evidence on competencies and risks (including cognitive off-loading) offers a useful conceptual foundation, and the explicit dual-track model with implementation and assessment recommendations provides practical value for educators developing curricula in AI-powered discovery.

major comments (2)

[Abstract and dual-track curriculum model section] Abstract and the section proposing the dual-track curriculum model: The central prescriptive claim that the workflow-aligned model 'must' replace tool-access approaches because it will produce improved scientific judgment, comparable learning gains across subgroups, and reduced cognitive off-loading is load-bearing for the thesis, yet the manuscript contains no original empirical data, pre/post assessments, controlled comparisons, or pilot results to substantiate these outcomes. It relies on synthesized external evidence without demonstrating the model's effectiveness.
[Outcome-oriented equity section] Section on outcome-oriented equity: The definition of equity as producing comparable gains in confidence calibration (alignment of student confidence with work quality), persistence, and research readiness is introduced as a key evaluation criterion, but the paper does not specify measurable indicators, assessment methods, or how institutions would implement and verify these across subgroups, leaving the equity claim without operational support.

minor comments (2)

[Abstract] The parenthetical definition of confidence calibration in the abstract ('defined as the alignment with students confidence and the quality or correctness of their work') is awkwardly phrased and could be clarified for precision and readability.
[Risks identification section] The manuscript would benefit from additional specific citations to empirical studies on cognitive off-loading in AI-assisted scientific workflows to strengthen the risk identification section.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback on our position paper. We address each major comment below, clarifying the scope of this work as a synthesis-driven proposal while committing to revisions that strengthen its framing and operational details.

read point-by-point responses

Referee: [Abstract and dual-track curriculum model section] Abstract and the section proposing the dual-track curriculum model: The central prescriptive claim that the workflow-aligned model 'must' replace tool-access approaches because it will produce improved scientific judgment, comparable learning gains across subgroups, and reduced cognitive off-loading is load-bearing for the thesis, yet the manuscript contains no original empirical data, pre/post assessments, controlled comparisons, or pilot results to substantiate these outcomes. It relies on synthesized external evidence without demonstrating the model's effectiveness.

Authors: This manuscript is a workshop-informed white paper and position article whose purpose is to synthesize external evidence on AI-related risks (such as cognitive off-loading), materials-informatics competencies, and equity considerations, then propose a workflow-aligned framework and dual-track curriculum model. It does not claim to present original empirical data or controlled evaluations, nor does it assert that the proposed outcomes have been demonstrated within this work. The prescriptive language is offered as a recommendation grounded in the cited literature rather than as a proven result. We agree that the load-bearing nature of the central claim warrants clearer framing. In revision we will update the abstract and dual-track section to explicitly describe the manuscript as a hypothesis-generating proposal that identifies risks and advocates for future empirical validation, pilot implementations, and comparative studies. This preserves the core argument while removing any implication of demonstrated effectiveness. revision: partial
Referee: [Outcome-oriented equity section] Section on outcome-oriented equity: The definition of equity as producing comparable gains in confidence calibration (alignment of student confidence with work quality), persistence, and research readiness is introduced as a key evaluation criterion, but the paper does not specify measurable indicators, assessment methods, or how institutions would implement and verify these across subgroups, leaving the equity claim without operational support.

Authors: The equity section introduces outcome-oriented equity as an evaluation criterion and references an assessment plan among the implementation recommendations. We acknowledge that greater specificity on indicators and verification methods would improve operational utility. In the revised manuscript we will expand the section to include example measurable indicators (e.g., pre/post use of validated confidence-calibration scales aligned with task performance, disaggregated persistence metrics such as course completion and research involvement rates, and rubric-based assessments of research readiness) together with practical implementation guidance such as subgroup data analysis protocols and iterative curriculum feedback mechanisms. revision: yes

Circularity Check

0 steps flagged

No circularity: conceptual position paper with no derivations or self-referential predictions.

full rationale

The manuscript is a workshop-informed white paper proposing a workflow-aligned AI literacy framework for materials science education. It contains no equations, fitted parameters, predictions, or derivation chains that could reduce to inputs by construction. The argument synthesizes external literature on AI risks and competencies, then offers curriculum recommendations and equity metrics as prescriptive guidance rather than derived results. No self-citation load-bearing steps, ansatz smuggling, or renaming of known results occur; the central claims rest on synthesized evidence and untested hypotheses about learning outcomes, which the paper itself does not claim to validate empirically within the text.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on domain assumptions about AI's role in materials science and the benefits of workflow alignment, presented without new empirical grounding or independent validation.

axioms (2)

domain assumption AI models increasingly support property prediction, experiment prioritization, and hypothesis generation in materials science
Stated as background in the opening of the abstract.
domain assumption The limiting factor is whether students and educators can use AI with domain-specific scientific judgment
Presented as the key premise driving the need for the framework.

pith-pipeline@v0.9.0 · 5559 in / 1424 out tokens · 96096 ms · 2026-05-12T03:23:42.667058+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We connect AI literacy to materials-informatics competencies: data provenance, domain-specific featurization, model validation, uncertainty quantification, physics informed reasoning, reproducibility, and experimental feedback.
IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

The central goal is to prepare students to become better scientists, not merely more efficient users of AI tools.

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

45 extracted references · 45 canonical work pages

[1]

Workshop for AI-powered materials discovery at Great Plains

University of South Dakota. Workshop for AI-powered materials discovery at Great Plains. Workshop webpage, Indico: https://aimaterialsworkshop.org/event/1/,

work page
[2]

Accessed April 30, 2026

Held June 22–25, 2025, University of South Dakota. Accessed April 30, 2026

work page 2025
[3]

AI competency framework for students

Fengchun Miao, Kelly Shiohira, and Natalie Lao. AI competency framework for students. Technical report, UNESCO, Paris, France, 2024

work page 2024
[4]

Kassorla, M

M. Kassorla, M. Georgieva, and A. Papini. AI literacy in teaching and learning: A durable framework for higher education. EDUCAUSE, October 2024

work page 2024
[5]

Mills, J

Kelly Mills, Pati Ruiz, Keun-woo Lee, Merijke Coenraad, Judi Fusco, Jeremy Roschelle, and Josh Weisgrau. AI literacy: A framework to understand, evaluate, and use emerging technology. Technical report, Digital Promise, May 2024. DOI: 10.51388/20.500.12265/218. 19

work page doi:10.51388/20.500.12265/218 2024
[6]

T. J. Oweida, A. Mahmood, M. D. Manning, S. Rigin, and Y. G. Yingling. Merging materials and data science: Opportunities, challenges, and education in materials informatics.MRS Advances, 5:329–346, 2020

work page 2020
[7]

A. Y.-T. Wang, R. J. Murdock, S. K. Kauwe, et al. Machine learning for materials scientists: An introductory guide toward best practices.Chemistry of Materials, 32:4954– 4965, 2020

work page 2020
[8]

B. L. DeCost, J. R. Hattrick-Simpers, Z. Trautt, A. G. Kusne, E. Campo, and M. L. Green. Scientific AI in materials science: A path to a sustainable and scalable paradigm. Machine Learning: Science and Technology, 1(3):033001, 2020

work page 2020
[9]

Boyce, R

B. Boyce, R. Dingreville, S. Desai, et al. Machine learning for materials science: Barriers to broader adoption.Matter, 6:1320–1323, 2023

work page 2023
[10]

AI in Education Report: New Cengage Group data shows growing GenAI adoption in K–12 and higher education

Cengage Group. AI in Education Report: New Cengage Group data shows growing GenAI adoption in K–12 and higher education. Press release and survey report, April

work page
[11]

Survey report on student and faculty use of generative AI

work page
[12]

DEC AI literacy framework, 2025

Digital Education Council. DEC AI literacy framework, 2025

work page 2025
[13]

Empowering learners for the age of AI: An AI literacy framework for primary and secondary education (review draft)

OECD and European Commission. Empowering learners for the age of AI: An AI literacy framework for primary and secondary education (review draft). Technical report, OECD and European Commission, May 2025

work page 2025
[14]

State AI guidance for K–12 schools

Daniel Kosta. State AI guidance for K–12 schools. Online resource, AI for Education, October 2025. Originally posted January 21, 2025; updated October 28, 2025

work page 2025
[15]

Casal-Otero, A

L. Casal-Otero, A. Catala, C. Fernandez-Morante, M. Taboada, B. Cebreiro, and S. Barro. AI literacy in K–12: A systematic literature review.International Journal of STEM Education, 10(1):29, 2023

work page 2023
[16]

Riding the AI wave: What’s happening in K–12 education? Center for Security and Emerging Technology, 2024

Ali Crawford and Cherry Wu. Riding the AI wave: What’s happening in K–12 education? Center for Security and Emerging Technology, 2024

work page 2024
[17]

Developing a model for AI across the curriculum: Transforming the higher education landscape via innovation in AI literacy

Jane Southworth, Kati Migliaccio, Joe Glover, Ja’Net Glover, David Reed, Christopher McCarty, Joel Brendemuhl, and Aaron Thomas. Developing a model for AI across the curriculum: Transforming the higher education landscape via innovation in AI literacy. Computers and Education: Artificial Intelligence, 4:100127, 2023

work page 2023
[18]

AI in computer science education: Closing the new digital divide in K–12

Nancy Mann Jackson. AI in computer science education: Closing the new digital divide in K–12. EdTech Magazine, November 2025

work page 2025
[19]

Balachandran, Dezhen Xue, and Ruihao Yuan

Turab Lookman, Prasanna V. Balachandran, Dezhen Xue, and Ruihao Yuan. Active learning in materials science with emphasis on adaptive sampling using uncertainties for targeted design.npj Computational Materials, 5(1):21, 2019

work page 2019
[20]

Mannodi-Kanakkithodi, A

A. Mannodi-Kanakkithodi, A. McDannald, S. Sun, et al. A framework for materials informatics education through workshops.MRS Bulletin, 48:560–569, 2023. 20

work page 2023
[21]

The Current State of Artificial Intelligence in Education

National Education Association. Report of the NEA task force on artificial intelligence in education. Technical report, National Education Association, June 2024. Includes section “The Current State of Artificial Intelligence in Education”; last updated October 22, 2024

work page 2024
[22]

L. Langreo. Teachers desperately need AI training: How many are getting it? Education Week, March 2024

work page 2024
[23]

Q. Tan. Reimagining teacher development in the era of generative AI: A scoping review. Teaching and Teacher Education, 168:105236, 2025

work page 2025
[24]

AI book club: An innovative professional development model for AI education

Irene Lee, Helen Zhang, Kate Moore, Xiaofei Zhou, Beatriz Perret, Yihong Cheng, Ruiying Zheng, and Grace Pu. AI book club: An innovative professional development model for AI education. InProceedings of the 53rd ACM Technical Symposium on Computer Science Education, pages 202–208, 2022

work page 2022
[25]

Zhang, I

H. Zhang, I. Lee, and K. Moore. Preparing teachers to teach artificial intelligence in classrooms: An exploratory study. InProceedings of the 17th International Conference of the Learning Sciences (ICLS 2023), pages 974–977, 2023

work page 2023
[26]

Zhang, I

H. Zhang, I. Lee, and K. Moore. An effectiveness study of teacher-led AI literacy curriculum in K–12 classrooms. InProceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 23318–23325, 2024

work page 2024
[27]

K. VanLehn. The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systems.Educational Psychologist, 46(4):197–221, 2011

work page 2011
[28]

J. A. Kulik and J. D. Fletcher. Effectiveness of intelligent tutoring systems: A meta- analytic review.Review of Educational Research, 86(1):42–78, 2016

work page 2016
[29]

Michelene T. H. Chi and Ruth Wylie. The ICAP framework: Linking cognitive engage- ment to active learning outcomes.Educational Psychologist, 49(4):219–243, 2014

work page 2014
[30]

Butler, Daniel W

Keith T. Butler, Daniel W. Davies, Hugh Cartwright, Olexandr Isayev, and Aron Walsh. Machine learning for molecular and materials science.Nature, 559(7715):547–555, 2018

work page 2018
[31]

Generative AI without guardrails can harm learning: Evidence from high school mathematics

Hamsa Bastani, Osbert Bastani, Alp Sungu, Haosen Ge, Ozge Kabakci, and Rei Mari- man. Generative AI without guardrails can harm learning: Evidence from high school mathematics. SSRN working paper, 2024. Working-paper version of the study later published inProceedings of the National Academy of Sciences

work page 2024
[32]

Beware of metacognitive laziness: Effects of generative artificial intelligence on learning motivation, processes, and performance

Yizhou Fan, Luzhen Tang, Huixiao Le, Kejie Shen, Shufang Tan, Yueying Zhao, Yuan Shen, Xinyu Li, and Dragan Gaˇ sevi´ c. Beware of metacognitive laziness: Effects of generative artificial intelligence on learning motivation, processes, and performance. British Journal of Educational Technology, 56(2):489–530, 2025

work page 2025
[33]

B. Z. Larson, C. Moser, A. Caza, K. Muehlfeld, and L. A. Colombo. Critical thinking in the age of generative AI.Academy of Management Learning & Education, 23(3):373–378, 2024. 21

work page 2024
[34]

Shaw and Gideon Nave

Steven D. Shaw and Gideon Nave. Thinking—fast, slow, and artificial: How AI is reshaping human reasoning and the rise of cognitive surrender. SSRN Working Paper 6097646, The Wharton School of the University of Pennsylvania, 2026. Version 20260111. DOI: 10.2139/ssrn.6097646

work page doi:10.2139/ssrn.6097646 2026
[35]

T. Lai, C. Xie, M. Ruan, Z. Wang, H. Lu, and S. Fu. Influence of artificial intelligence in education on adolescents’ social adaptability: The mediatory role of social support. PLOS ONE, 18(3):e0283170, 2023

work page 2023
[36]

Guidance for generative AI in education and research

Fengchun Miao and Wayne Holmes. Guidance for generative AI in education and research. Technical report, UNESCO, Paris, France, 2023. Policy guidance document

work page 2023
[37]

Kaufman, Ashley Woo, Joshua Eagan, Sabrina Lee, and Emma B

Julia H. Kaufman, Ashley Woo, Joshua Eagan, Sabrina Lee, and Emma B. Kassan. Uneven adoption of artificial intelligence tools among U.S. teachers and principals in the 2023–2024 school year. Technical Report RR-A134-25, RAND Corporation, Santa Monica, CA, 2025

work page 2023
[38]

H. Pham, T. Kohli, E. Olick Llano, I. Nokuri, and A. Weinstock. How will AI impact racial disparities in education. Stanford Center for Racial Justice, 2024

work page 2024
[39]

Does ChatGPT enhance student learning? a systematic review and meta-analysis of experimental studies

Ruiqi Deng, Maoli Jiang, Xinlu Yu, Yuyan Lu, and Shasha Liu. Does ChatGPT enhance student learning? a systematic review and meta-analysis of experimental studies. Computers & Education, 227:105224, 2025

work page 2025
[40]

Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant for essay writing task

Nataliya Kosmyna, Eugene Hauptmann, Ye Tong Yuan, Jessica Situ, Xian-Hao Liao, Ashly Vivian Beresnitzky, Iris Braunstein, and Pattie Maes. Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant for essay writing task. arXiv preprint arXiv:2506.08872, 2025

work page arXiv 2025
[41]

Lehmann, P

M. Lehmann, P. B. Cornelius, and F. J. Sting. AI meets the classroom: When does ChatGPT harm learning? arXiv preprint arXiv:2409.09047, 2024

work page arXiv 2024
[42]

Bastani, O

H. Bastani, O. Bastani, A. Sungu, H. Ge, O. Kabakci, and R. Mariman. Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122(26):e2422633122, 2025

work page 2025
[43]

Is ChatGPT a boon or a bane for learning? experimental evidence across task formats and chatbot designs

Andy Tao Li, De Liu, and Teng Ye. Is ChatGPT a boon or a bane for learning? experimental evidence across task formats and chatbot designs. SSRN working paper, September 2025. Date written September 13, 2025; posted September 27, 2025

work page 2025
[44]

Civic education in the age of AI: Should we trust AI-generated lesson plans?Contemporary Issues in Technology and Teacher Education, 25(3), 2025

Torrey Trust, Robert Maloy, Chenyang Xu, and Kael Pelletier. Civic education in the age of AI: Should we trust AI-generated lesson plans?Contemporary Issues in Technology and Teacher Education, 25(3), 2025

work page 2025
[45]

Veldhuis, P

A. Veldhuis, P. Y. Lo, S. Kenny, and A. N. Antle. Critical artificial intelligence literacy: A scoping review and framework synthesis.International Journal of Child-Computer Interaction, 43:100708, 2025. 22

work page 2025