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-2026-2710
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-11-12T17:12:53.598347Z
status: published
lens: g_space
sector: education
headline: Artificial intelligence-enhanced inquiry-based learning and higher-order thinking skills: a meta-analysis
dek: This meta-analysis synthesizes empirical evidence from 17 studies (N = 2,052) examining the effectiveness of artificial intelligence-enhanced inquiry-based learning on higher-order thinking skills (HOTS). A random-effects model revealed a medium-to-large positive effect (g = 0.696, 95% CI [0.607, 0.785], p < 0.001). Subgroup analyses showed that generative AI technologies, particularly ChatGPT (g = 0.754), outperformed traditional intelligent tutoring systems (g = 0.614), and problem-based learning demonstrated…
gain_title: A meta-analysis of 17 studies with 2,052 students found AI-enhanced inquiry-based learning improved higher-order thinking skills, with a medium-to-large pooled effect (g = 0.696).
problem_title: (none)
trace_subject: (none)
gain_reading: A meta-analysis of 17 studies with 2,052 students found AI-enhanced inquiry-based learning improved higher-order thinking skills, with a medium-to-large pooled effect (g = 0.696).
gain_evidence: A random-effects model revealed a medium-to-large positive effect (g = 0.696, 95% CI [0.607, 0.785], p < 0.001).
problem_reading: (none)
problem_evidence: (none)
quick_read: A meta-analysis pooling 17 studies and 2,052 participants found that inquiry-based learning supported by artificial intelligence improved higher-order thinking skills, with a pooled effect of g = 0.696 (95% CI 0.607 to 0.785).

Generative AI tools did better than older intelligent tutoring systems: ChatGPT-based interventions reached g = 0.754 against g = 0.614 for tutoring systems. Problem-based learning showed the largest effect, at g = 0.812.

Longer interventions were associated with larger effects in a meta-regression. Risk-of-bias ratings classed 59% of the studies as low risk, and the authors report that sensitivity analyses held up and publication bias was minimal.
limitation: Only 59% of the included studies were rated low risk of bias under Cochrane Risk of Bias 2.0, so the remainder carry greater risk of bias.
tag: Evidence-backed gain
key_points: The analysis pooled 17 studies covering 2,052 participants using a random-effects model. | The overall effect on higher-order thinking skills was g = 0.696 (95% CI 0.607 to 0.785, p < 0.001). | Generative AI, particularly ChatGPT, produced g = 0.754, compared with g = 0.614 for traditional intelligent tutoring systems. | Problem-based learning had the highest effect among the approaches compared, at g = 0.812. | Intervention duration was a significant predictor of effect size in the meta-regression. | Cochrane Risk of Bias 2.0 rated 59% of included studies as low risk.
rundown: The review synthesised empirical studies on AI-enhanced inquiry-based learning and its effect on higher-order thinking skills, using a random-effects model across 17 studies and 2,052 participants. The pooled result was a medium-to-large positive effect.

Subgroup analyses compared technologies and teaching approaches. Generative AI tools outperformed traditional intelligent tutoring systems, and ChatGPT specifically reached g = 0.754. Problem-based learning produced the largest effect at g = 0.812.

Meta-regression pointed to intervention duration as a significant predictor of effect size. For quality, Cochrane Risk of Bias 2.0 classed 59% of the studies as low risk, and sensitivity analyses and publication-bias checks supported the findings.
sources:
- peer_reviewed | Learning: Research and Practice | https://doi.org/10.1080/23735082.2026.2672411 | 2026-05-22
prev: 0000000000000000000000000000000000000000000000000000000000000000

SHA-256 hash

745eb94c234e90e3e9aced0beb021efb7877275835c96e72e1804fed3a3d73d2

Previous version’s hash

0000000000000000000000000000000000000000000000000000000000000000