Truace
Subscribe
70/100

Moderate evidence

ProblemPublished recordHealthProblem Space

Full story
More
Export record JSONEmailLinkedInX

Is it time to introduce an artificial intelligence curriculum in undergraduate medical education? medical students’ perspectives: a systematic review and meta-analysis

Truvace record
TRV-2026-2784raw data (JSON)
Version
v1Published
First published
Evidence type
Systematic review or meta-analysis
Sources
1 sourcePeer-reviewed
sha256
0405f1c841a6…check it

BACKGROUND: Although artificial intelligence (AI) is increasingly transforming healthcare systems, its systematic integration into undergraduate medical education (UME) remains limited. Despite widespread recognition of AI's potential to enhance clinical practice, AI-related competencies are still inadequately embedded in most medical curricula, even though medical students generally express positive attitudes toward AI. OBJECTIVES: This systematic review and meta-analysis synthesizes the current literature on m…

The problem

A systematic review of 26 studies of 20,963 medical students found 83.1% agreed AI training is necessary, yet only 36.4% felt confident applying AI in clinical practice.

What this doesn’t fix

The meta-analysis found extreme heterogeneity across studies (I² = 98.5%), so the pooled proportions mask large differences between the underlying studies.

Sources

1 on file
  1. BMC Medical Educationdoi:10.1186/s12909-026-09630-9
    Peer-reviewed2026-06-17

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.

  • Medical students with positive attitudes toward AI curriculum, adjusted for possible publication bias (trim-and-fill)

    Proportion 0.818 (95% CI 0.771 to 0.861)

    26 studies · 20,963 people · search to February 2025

    The review’s own words
    However, the use of the Trim-and-Fill method gave the imputation of six studies, shifting the random-effects pooled proportion to 0.818 (95% CI: 0.771–0.861) with a prediction interval of [0.500; 0.991] (see Supplementary Material: Appendix 3, Figure S1).

  • Positive attitudes toward AI curriculum among medical students in the Middle East

    Proportion 0.730 (95% CI 0.660 to 0.800)

    search to February 2025

    The review’s own words
    The pooled proportion with favorable attitudes toward AI was: North America (0.780, 95% CI: 0.660–0.880), Asia (0.780, 95% CI: 0.700–0.860), the Middle East (0.730, 95% CI: 0.660–0.800), and other regions (0.910, 95% CI: 0.830–0.960).

  • Positive attitudes toward AI curriculum among medical students in North America

    Proportion 0.780 (95% CI 0.660 to 0.880)

    search to February 2025

    The review’s own words
    The pooled proportion with favorable attitudes toward AI was: North America (0.780, 95% CI: 0.660–0.880), Asia (0.780, 95% CI: 0.700–0.860), the Middle East (0.730, 95% CI: 0.660–0.800), and other regions (0.910, 95% CI: 0.830–0.960).

  • Medical students with positive attitudes toward AI curriculum integration (common-effect model)

    Proportion 0.779 (95% CI 0.774 to 0.783)

    26 studies · 20,963 people · search to February 2025

    The review’s own words
    Overall, the proportion with favorable attitudes was 0.779 (95% CI: 0.774–0.783) under the common-effect model and 0.780 (95% CI: 0.730–0.820) under the random-effects model using the DerSimonian-Laird estimator (Fig. 2 ).

  • Positive attitudes toward AI curriculum among medical students in other regions

    Proportion 0.910 (95% CI 0.830 to 0.960)

    search to February 2025

    The review’s own words
    The pooled proportion with favorable attitudes toward AI was: North America (0.780, 95% CI: 0.660–0.880), Asia (0.780, 95% CI: 0.700–0.860), the Middle East (0.730, 95% CI: 0.660–0.800), and other regions (0.910, 95% CI: 0.830–0.960).

  • Positive attitudes toward AI curriculum among medical students in Asia

    Proportion 0.780 (95% CI 0.700 to 0.860)

    search to February 2025

    The review’s own words
    The pooled proportion with favorable attitudes toward AI was: North America (0.780, 95% CI: 0.660–0.880), Asia (0.780, 95% CI: 0.700–0.860), the Middle East (0.730, 95% CI: 0.660–0.800), and other regions (0.910, 95% CI: 0.830–0.960).

  • Medical students with positive attitudes toward AI curriculum integration (random-effects)

    Proportion 0.780 (95% CI 0.730 to 0.820)

    26 studies · 20,963 people · search to February 2025

    The review’s own words
    Overall, the proportion with favorable attitudes was 0.779 (95% CI: 0.774–0.783) under the common-effect model and 0.780 (95% CI: 0.730–0.820) under the random-effects model using the DerSimonian-Laird estimator (Fig. 2 ).

  • Positive attitudes toward AI curriculum in studies using standardized questionnaires

    Proportion 0.780 (95% CI 0.690 to 0.850)

    search to February 2025

    The review’s own words
    Stratified by type of questionnaire, researcher-developed surveys gave a proportion of 0.780 (95% CI: 0.710–0.840), and standardized tools gave 0.780 (95% CI: 0.690–0.850).

  • Positive attitudes toward AI curriculum in studies using researcher-developed surveys

    Proportion 0.780 (95% CI 0.710 to 0.840)

    search to February 2025

    The review’s own words
    Stratified by type of questionnaire, researcher-developed surveys gave a proportion of 0.780 (95% CI: 0.710–0.840), and standardized tools gave 0.780 (95% CI: 0.690–0.850).

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

Cite this record

v1

Truvace Impact Record TRV-2026-2784, v1: “Is it time to introduce an artificial intelligence curriculum in undergraduate medical education? medical students’ perspectives: a systematic review and meta-analysis.” Truvace, 2026-11-12. /record/TRV-2026-2784 (accessed on the day you copy it). sha256 0405f1c841a67601…

Change history

1 version

Every change made to this record since it was first published, and why. Nothing has changed since it went up.

  1. Publishedv10405f1c841a6…

    Certified into the record

Check this record hasn’t been altered

Check this record hasn’t been altered
How to check it yourself, without trusting this page

Download any version from /api/record/TRV-2026-2784 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.