Center for Practical AI

CPAI Issue Brief · Education & AI literacy

Safe AI Use

The same AI tool produces opposite outcomes depending on how it is used — and usage guidance is largely absent while adoption is near-universal.

What’s happening

AI assistants are now in daily use across work, school, and home. The evidence increasingly shows that the outcome — whether AI use builds capability or erodes it — depends less on how much people use it than on how.

What the evidence shows

In a randomized trial of about 1,000 students, unrestricted GPT-4 raised practice performance 48% but left those students 17% worse on a later unassisted exam; a version redesigned to give hints instead of answers eliminated the harm. A separate set of randomized studies found that people who used AI mainly for direct answers underperformed afterward, while those who used it for hints showed no deficit. One positive result, one negative — both randomized. The brief's credibility rests on presenting both.

Well-designed AI tutoring can roughly double learning gains over conventional instruction. The common thread across the evidence is design and usage mode, not access alone.

+48% / −17%

unrestricted AI raised practice but lowered unassisted exam scores

Bastani et al. 2025, PNAS — RCT

61% / 27%

used AI mainly for answers vs. for hints — only the answer-seekers underperformed

Liu et al. 2026 — RCT subgroup

Where it reaches constituents

Students, workers, and families are adopting AI universally while usage guidance lags. Roughly three in four knowledge workers already use AI at work, and most received no training — so the habits that determine whether AI helps or harms are being formed by default, not design.

The current legal & regulatory landscape

The EU AI Act's Article 4 AI-literacy obligation has been in force since February 2025 — the first legal AI-literacy mandate. There is no comparable federal AI-literacy requirement in the United States. North Carolina's AI Strategic Roadmap commits to foundational AI-literacy training across all 100 counties by 2028.

Considerations policymakers are weighing

  • ·How AI-literacy programs define "literacy" — tool mechanics versus the usage habits and evaluation skills that the evidence links to outcomes.
  • ·Whether learning products disclose whether they are designed to give hints or answers.
  • ·How programs measure outcomes (durable capability) rather than attendance.

Listed as live debates, not recommendations. CPAI does not take a position on how these should be resolved.

This brief condenses a full, sourced public guide. The complete evidence and citations:

Healthy AI Use (5-guide series) →

Key sources

Official policy / primary sourceEuropean Union (2025) — EU AI Act, Article 4 — AI literacy obligation ↗In force since February 2, 2025 — the first legal AI-literacy mandate, requiring providers and deployers to ensure staff have a sufficient level of AI literacy. (A November 2025 Digital Omnibus proposal may soften the general obligation.) Included as international context.
Official policy / primary sourceState of North Carolina (2026) — Statewide AI Strategic Roadmap ↗Issued July 1, 2026 under Executive Order 24 (Sept 2, 2025). Sets 17 goals across Protect / Prepare / Transform with targets through December 2028 — including foundational AI-literacy training in all 100 counties (via NCWorks, community colleges, and libraries) and credentialing 50,000+ residents in AI skills. These are commitments with deadlines, not appropriations.
Company-reportedMicrosoft & LinkedIn (2024) — Work Trend Index — AI at Work Is Here ↗Company-reported survey, 31,000 workers across 31 countries: 75% of knowledge workers already use generative AI at work and 78% bring their own tools without employer guidance, yet only 39% of AI users had received company training. Vendor-reported — read as adoption signal, not independent measurement.
Randomized controlled trialBastani, Bastani, Sungu, Ge, Kabakcı & Mariman (2025) — Generative AI Can Harm Learning ↗PNAS. Randomized trial, ~1,000 students across ~50 Turkish high-school math classes. Unrestricted GPT-4 access raised practice performance 48% but lowered the later unassisted exam 17% versus never-AI controls. A teacher-designed, hints-only tutor version raised practice performance 127% — and eliminated the exam harm.
Randomized trial · preprint, not yet peer-reviewedLiu et al. (2026), usage-mode analysis — Hint-seekers vs. answer-seekers within the persistence trials ↗Within the second persistence experiment, 61% of participants said they used AI mainly for direct answers, 27% for hints and clarification. The groups were indistinguishable before AI use — but afterward, answer-seekers underperformed the no-AI control (d=0.36) and skipped more, while hint-seekers showed no deficit at all. Mode was self-chosen, so disposition and mode cannot be separated.
Randomized controlled trialKestin, Miller, Klales, Milbourne & Ponti (2025) — AI Tutoring Outperforms In-Class Active Learning ↗Scientific Reports. Crossover randomized trial, ~190 Harvard intro-physics students. A GPT-4 tutor engineered to work one step at a time, never reveal full solutions, and encourage attempts first produced roughly double the learning gains of in-class active learning, in less time. Two lessons, elite sample, carefully engineered tutor — not raw ChatGPT.
Randomized trial · preprint, not yet peer-reviewedLiu, Christian, Dumbalska, Bakker & Dubey (2026) — AI Assistance Reduces Persistence and Hurts Independent Performance ↗Three randomized studies, N=1,222. After roughly ten minutes of AI-assisted work, people solved fewer problems unaided (89% → 76%) and skipped more without attempting them at all (1% → 8%).

A nonpartisan resource

The Center for Practical AI is a nonpartisan 501(c)(3) nonprofit. We provide education, research, and analysis, and we offer briefings and testimony on request. We do not endorse candidates or lobby for or against specific legislation. Everything here describes the evidence and the current landscape — the policy choices are yours.

Want CPAI to brief your office on this?

We provide nonpartisan briefings, research summaries, and testimony on request.

Request a briefing →