aceModerate evidence
More
Review of 33 studies finds AI tools gave university students a learning boost
Study title: Generative AI in higher education: a meta-analysis of intellectual and social–emotional outcomes
- Evidence type
- Systematic review or meta-analysis
- Sample
- n = 3,394
- Sources
- 1 sourcePeer-reviewed
- Published
- 2 min read
- Truvace record
- TRV-2026-2783cite or check it
Study title: Generative AI in higher education: a meta-analysis of intellectual and social–emotional outcomes
A systematic review and meta-analysis pooled 33 experimental and quasi-experimental studies covering 3,394 higher education students. It found generative AI had a significant positive effect on intellectual outcomes (Hedges' g = 1.096) and a smaller one on social-emotional outcomes (g = 0.301).
Read the study abstract
Background Generative AI has rapidly entered higher education, yet evidence on its effects across learning domains remains fragmented. This study synthesized experimental evidence on whether generative AI improves intellectual outcomes and social emotional outcomes among higher education students. Methods A systematic review and meta-analysis was conducted on 33 experimental and quasi-experimental studies involving 3,394 participants. The included studies were published between 2021 and 2024, and the literature…
The gain
A meta-analysis of 33 experimental studies with 3,394 university students found generative AI raised intellectual outcomes (Hedges' g = 1.096) and social-emotional outcomes (g = 0.301) compared with control conditions.
In brief
The intellectual-outcome estimate has a very wide confidence interval (0.087 to 2.104), and the gap between the two outcome domains was not statistically significant (p = 0.252). The authors call generative AI a promising resource, particularly for intellectual outcomes.
Confidence in the intellectual result is weaker: publication bias was detected for those outcomes and the results were sensitive to individual studies. The social-emotional findings were relatively stable.
Main points
- The review pooled 33 experimental and quasi-experimental studies with 3,394 participants, published between 2021 and 2024.
- Intellectual outcomes showed a pooled effect of Hedges' g = 1.096 (95% CI 0.087 to 2.104, p = 0.033).
- Social-emotional outcomes showed a smaller pooled effect of g = 0.301 (95% CI 0.048 to 0.553, p = 0.020).
- The difference between the two outcome domains was not statistically significant (coefficient = -0.739, p = 0.252).
- Publication bias was detected for intellectual outcomes but not for social-emotional outcomes.
- Functional type, intervention duration and knowledge domain did not significantly explain differences between studies.
What the review found when it combined the studies
These are the review’s own pooled results, copied from its abstract. Truvace did not recalculate them.
Social-emotional outcomes of higher education students using generative AI
Hedges' g 0.301 (95% CI 0.048 to 0.553)
12 studies · 3,394 people · search to January 26, 2026
The review’s own words
a smaller but significant positive effect on social emotional outcomes (Hedges’ g = 0.301, 95% CI 0.048 to 0.553, p = 0.020)
Intellectual outcomes of higher education students using generative AI
Hedges' g 1.096 (95% CI 0.087 to 2.104)
21 studies · 3,394 people · search to January 26, 2026
The review’s own words
Generative AI demonstrated a significant positive effect on intellectual outcomes (Hedges’ g = 1.096, 95% CI 0.087 to 2.104, p = 0.033)
Code checks every number above against the review’s text before it is shown. The short labels are an AI model’s reading of that text, so the review’s own sentence sits under each one. A pooled figure is only as strong as the studies behind it. How Truvace uses pooled results
The full story
The authors searched databases from inception to January 26, 2026, and included experimental and quasi-experimental studies published between 2021 and 2024. They used random-effects models in Stata 18 and registered the review in PROSPERO (CRD420261361622).
Beyond the pooled effects, the team ran publication bias, sensitivity and moderator analyses. Social-emotional findings held up relatively well, while intellectual outcomes shifted depending on which individual studies were included.
Moderator analyses found that the type of AI function, how long the intervention lasted and the subject area did not significantly explain differences between studies at the omnibus level. The authors describe the evidence base as heterogeneous and methodologically uneven.
Sources
1 on file- Frontiers in Psychologydoi:10.3389/fpsyg.2026.1848745
Talk it over
Who gets this benefit, and who was left out of the study or story?
Review of 33 studies finds AI tools gave university students a learning boosthttps://truvace.com/item/generative-ai-in-higher-education-a-meta-analysis-of-intellectual-and-social-emo
Who gets this benefit, and who was left out of the study or story?
The gain A meta-analysis of 33 experimental studies with 3,394 university students found generative AI raised intellectual outcomes (Hedges' g = 1.096) and social-emotional outcomes (g = 0.301) compared with control conditions.
Answer in the threadHow do we know? What can this kind of evidence show, and what can’t it?
The evidence This rests on a systematic review or meta-analysis with a sample of 3,394, from Frontiers in Psychology.
Answer in the threadWhat does it leave open? What else would you want to know before acting on it?
The record’s own caveat The authors say the evidence base is heterogeneous and methodologically uneven, publication bias was detected for intellectual outcomes, and those results were sensitive to individual studies.
Answer in the threadHave you seen AI used this way where you study, work or live? Did it go the same way?
Answer in the thread
The questions are the same on every record. The details under them are taken from this record as written.
What readers say