Generative AI without guardrails can harm learning: Evidence from high school mathematics
Proceedings of the National Academy of Sciences · 2025-07 · 122 (26)
DOI: 10.1073/pnas.2422633122 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the National Academy of Sciences, online erschienen am 25.06.2025. Verfasst von Hamsa Bastani und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Operations, Information, and Decisions, Wharton School, University of Pennsylvania, Philadelphia, PA 19104. Hauptthemen laut MEDLINE: Learning, Mathematics, Artificial Intelligence. Die Arbeit wurde bisher 372-mal zitiert (Crossref) und nennt 28 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Wharton AI and Analytics Initiative, Fishman Davidson Center at Wharton, Wharton Global Initiatives.
Abstract (Originalsprache)
Generative AI is poised to revolutionize how humans work, and has already demonstrated promise in significantly improving human productivity. A key question is how generative AI affects learning—namely, how humans acquire new skills as they perform tasks. Learning is critical to long-term productivity, especially since generative AI is fallible and users must check its outputs. We study this question via a field experiment where we provide nearly a thousand high school math students with access to generative AI tutors. To understand the differential impact of tool design on learning, we deploy two generative AI tutors: one that mimics a standard ChatGPT interface (“GPT Base”) and one with prompts designed to safeguard learning (“GPT Tutor”). Consistent with prior work, our results show that having GPT-4 access while solving problems significantly improves performance (48% improvement in grades for GPT Base and 127% for GPT Tutor). However, we additionally find that when access is subsequently taken away, students actually perform worse than those who never had access (17% reduction in grades for GPT Base)—i.e., unfettered access to GPT-4 can harm educational outcomes. These negative learning effects are largely mitigated by the safeguards in GPT Tutor. Without guardrails, students attempt to use GPT-4 as a “crutch” during practice problem sessions, and subsequently perform worse on their own. Thus, decision-makers must be cautious about design choices underlying generative AI deployments to preserve skill learning and long-term productivity.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Operations, Information, and Decisions, Wharton School, University of Pennsylvania, Philadelphia, PA 19104.
- Wharton AI & Analytics, Philadelphia, PA 19104.
- Department of Operations, Information, and Decisions, School of Engineering and Applied Science, University of Pennsylvania, Philadelphia, PA 19104.
- Department of Mathematics, Budapest British International School, Budapest 1125, Hungary.
- Independent, Philadelphia, PA 19104.
ORCID-Profile: Hamsa Bastani, Alp Sungu, Haosen Ge
Förderung
- Wharton AI and Analytics Initiative
- Fishman Davidson Center at Wharton
- Wharton Global Initiatives
Zitieren
APA
Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122(26). https://doi.org/10.1073/pnas.2422633122
Vancouver
Bastani H, Bastani O, Sungu A, Ge H, Kabakcı Ö, Mariman R. Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences. 2025;122(26). doi:10.1073/pnas.2422633122
BibTeX
@article{Bastani2025,
author = {Bastani, Hamsa and Bastani, Osbert and Sungu, Alp and Ge, Haosen and Kabakcı, Özge and Mariman, Rei},
title = {Generative AI without guardrails can harm learning: Evidence from high school mathematics},
journal = {Proceedings of the National Academy of Sciences},
year = {2025},
volume = {122},
number = {26},
doi = {10.1073/pnas.2422633122}
}
EUVIMED