Work

Experimental Evidence on the Learning Impact of Generative AI

Year: 2026

Type: preprint

Study designs: Randomized Controlled Trial, Clinical Trial

Abstract: We study whether access to off-the-shelf generative AI changes how much students learn in a timed assignment. In a proctored, in-person randomized experiment, undergraduates study an unfamiliar topic and write an essay with or without AI, then complete unaided assessments immediately and one week later. AI access raises test scores, and a smaller gain persists a week later. Gains remain for students who use AI as a tutor ("augmentation") and fade for students who have AI write for them ("automation"). We document two mechanisms: students shift time from writing toward reading and searching for information, and report greater learning enjoyment.

Source: arXiv (Cornell University)

DOI: 10.48550/arxiv.2607.08849

Authors: Contractor, Zara, Reyes, Germán

Institutions: Cornell University, The University of Texas Rio Grande Valley, Ohio University, University of Missouri, Universidad Nacional de La Plata

Language: English

Topic: Artificial Intelligence in Education

Subfield: Education

Field: Social Sciences

Domain: Social Sciences

SDG: Quality education

Open Access status: green

OpenAlex ID: W7168156755

Cites: 8

Cited by: 4

Related to: 10