The exact text we saved

This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.

TRUVACE RECORD VERSION
record: TRV-2026-2781
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-11-12T17:23:17.699742Z
status: published
lens: g_space
sector: education
headline: Impact of artificial intelligence integration in STEM education on learners’ academic performance: A meta-analytic review
dek: The integration of artificial intelligence (AI) technologies into science, technology, engineering, and mathematics (STEM) education has attracted growing scholarly interest due to its potential to improve instructional quality and academic outcomes. This study quantitatively evaluated the effect of AI integration in STEM education on learners’ academic achievement using a meta-analytic approach. Following PRISMA guidelines, empirical studies published between 2015 and 2026 were retrieved from Web of Science, IE…
gain_title: A meta-analysis of 15 controlled studies found students taught STEM subjects with AI tools scored higher than control groups, with a pooled effect of g = 1.57 (95% CI 0.985 to 2.156), shrinking to 0.89 after a publication-bias adjustment.
problem_title: (none)
trace_subject: (none)
gain_reading: A meta-analysis of 15 controlled studies found students taught STEM subjects with AI tools scored higher than control groups, with a pooled effect of g = 1.57 (95% CI 0.985 to 2.156), shrinking to 0.89 after a publication-bias adjustment.
gain_evidence: The results showed a statistically significant positive effect of AI-based STEM interventions on academic achievement (g = 1.57, 95% CI [0.985, 2.156], Z = 5.256, p < 0.001). | the pooled effect size decreased from 1.57 to 0.89 after adjustment for potentially missing studies
problem_reading: (none)
problem_evidence: (none)
quick_read: A meta-analysis of 15 experimental or quasi-experimental studies, published between 2015 and 2026, compared STEM students taught with AI tools against control groups. The pooled effect on academic achievement was g = 1.57 (95% CI 0.985 to 2.156, p < 0.001), which the authors call a large positive effect.

Results varied widely between studies: heterogeneity was substantial (Q = 72.467, I² = 73%), and the 95% prediction interval ran from -0.26 to 3.40. That means a new study in a comparable setting could show little or no benefit. Individual study effects ranged from 0.23 to 4.53.

Adjusting for possible missing studies cut the pooled effect from 1.57 to 0.89. The effect stayed positive, but the authors say some influence of publication bias cannot be ruled out. They conclude the evidence points to a generally positive association, not a uniform effect in every classroom.
limitation: The review included only 15 studies, no formal risk-of-bias assessment was done, heterogeneity was substantial (I² = 73%), and the trim-and-fill adjustment cut the pooled effect from 1.57 to 0.89.
tag: Evidence-backed gain
key_points: The review searched Web of Science, IEEE Xplore and Google Scholar for studies from 2015 to 2026 and followed PRISMA guidelines; 894 records were identified and 15 studies were included. | Only studies with an experimental or quasi-experimental design, a control group and enough statistics to calculate an effect size qualified. | The pooled random-effects estimate was g = 1.57 (95% CI 0.985 to 2.156, Z = 5.256, p < 0.001). | Heterogeneity was substantial (Q = 72.467, df = 14, I² = 73%), and the prediction interval of -0.26 to 3.40 crosses zero. | Individual study effects ranged from 0.23 (Thai et al., 2022) to 4.53 (Essel et al., 2022). | Trim-and-fill adjustment lowered the pooled effect from 1.57 to 0.89; Begg and Egger tests found no significant publication bias, and Rosenthal's fail-safe N was 790. | Moderator analysis of outcome type, tool type and instructional approach explained none of the between-study variance (R² analog = 0.00, p = 0.789).
rundown: The authors, from universities in Kazakhstan, pooled studies of intelligent tutoring systems, adaptive learning platforms, chatbots and generative AI tools used in STEM teaching. They calculated Hedges' g for each study and combined them with a random-effects model in Comprehensive Meta-Analysis 4.0.

Publication-bias checks gave mixed signals. Begg and Mazumdar (Kendall's tau = 0.000, p = 0.500) and Egger's test (intercept 0.987, p = 0.213) found no significant bias, and the fail-safe N was 790. Even so, trim-and-fill reduced the estimate to 0.89, and the authors say bias cannot be fully excluded.

A meta-regression tested outcome type, instructional tool and approach as explanations for the differences between studies. None was significant (p = 0.420, 0.872 and 0.427), and residual heterogeneity stayed at I² = 76.02%. The authors say factors such as educational level, learner characteristics, intervention duration and implementation quality still need testing.
sources:
- peer_reviewed | International Journal of ADVANCED AND APPLIED SCIENCES | https://doi.org/10.21833/ijaas.2026.07.020 | 2026-07-01
prev: 0000000000000000000000000000000000000000000000000000000000000000

SHA-256 hash

7341e5f003c8c1dd5db7e7d49037624ca171e7c2a5c76f45a715ab9a1d5c2f65

Previous version’s hash

0000000000000000000000000000000000000000000000000000000000000000