Sungu, Lira and Duckworth, “Generative AI Can Harm Teaching”

In a randomized field experiment across a chain of middle and high schools in Turkey, giving teachers a generative-AI support tool reduced students’ intrinsic motivation by 0.11 standard deviations. Average academic performance did not change, but students taught by lower-performing teachers experienced significant declines in both performance and confidence, showing that a tool that makes lesson preparation easier for teachers does not automatically improve the student experience.

AIImpactLab story signal

Analysis byAIImpactLab Editorial Desk

PublishedJuly 20, 2026

Reading time3 minutes

Why it matters

Most participating teachers used the tool to create lecture materials, syllabi, and exercises. The result exposes a principal-agent problem in AI-assisted education: the person receiving the productivity benefit is not necessarily the person bearing the learning cost.

The uneven performance effect also matters. A neutral average can conceal greater harm where teaching quality is already weaker, making implementation support and student-centered evaluation essential before broad deployment.

Primary trail

Go to the source

Read the evidence behind this analysis. External links open in a new tab.

SSRN — Generative AI Can Harm Teaching

Why it matters

Go to the source

Related impact

Cash et al., “Is AI making us stupid?”

Liu et al., “Integrating chemical priors and physical laws to mitigate hallucinations in structure-based drug design”

Tang et al., “Reinforcement learning for treatment decision-making in sepsis: a scoping review”