Student or educator? Sign up with your school email and get 6 months of Pro free.

The DOI matches its Crossref record; 74 of the 74 references it cites were found in a registry or an index; 10 of the 64 sources its claims cite were checked for support.

Multinational machine learning investigation of whether generative AI interactions augment or substitute cognitive skills and study efficiency in global higher education

Open access

Abstract

While GenAI has been rapidly integrated into higher education, little empirical evidence explains how specific usage patterns drive cognitive augmentation versus cognitive offloading remains limited, as most studies treat usage uniformly. This study fills that gap by examining how GenAI habits, metacognitive behaviour, and emotions jointly predict cognitive skills and study efficiency as a distinct outcome across a global context. A secondary analysis was conducted on a publicly available cross-sectional multinational survey dataset of N = 14,949 higher education students, drawn from an original sample of 22,963 respondents across 120 countries. A five-layer analytical framework was applied: (1) measurement model validation using Confirmatory Factor Analysis (CFA) and Fornell-Larcker criterion; (2) multivariate OLS regression was used to predict critical thinking gain; (3) benchmarking of four machine learning algorithm (Linear Regression, Random Forest, XGBoost and Gradient Boosting) was used on a stratified 70/30 train-test spilt; (4) SHapley Additive exPlanations (SHAP) analysis was used on best performing model; and (5) K-Mean clustering (k = 4) was used across the three dimensions space of Cognitive Skill, Study Efficiency, and GenAI Usage. Robustness was assessed through five-fold cross-validation, discipline-stratified and subgroup analysis. The best-performing machine learning models demonstrated moderate predictive performance, explaining 28.18% of the variance in Cognitive Skill and 30.00% of the variance in Study Efficiency. The SHAP analysis result shows that Self-testing Habits (SHAP = 0.128) was the single strongest predictor of cognitive skill development. GenAI Usage Frequency (SHAP = 0.184) was the dominant driver of Study Efficiency. Four empirically distinct students’ interaction profiles were identified: (1) Augmentor (high skill, high efficiency, highest GenAI use; n = 2900); (2) Optimiser (moderate balance; n = 5500); (3) Traditionalist (efficient with lower GenAI use; n = 3000); and Substitutor (disengaged, low performance; n = 2000). Institutional funding type was a significant predictor of gains in study efficiency. The study concluded that effective human-GenAI interaction is not a function of usage frequency alone but of interaction quality, task type, and metacognitive structure within which AI tools are embedded. The cognitive skill efficiency trade-off is real but avoidable: Augmentor-profile students demonstrate that high efficiency and high cognitive skill co-exist when GenAI use is intensive, structured, and metacognitively scaffolded. Higher education institutions must move beyond binary use/no-use policies toward evidence-based frameworks that promote generative, metacognitively embedded interaction to maximise cognitive augmentation and mitigate passive substitution.

Overview

CoChat Summary

TL;DR

This secondary analysis of a multinational survey (N = 14,949) finds that cognitive skill development in GenAI-assisted learning is driven mainly by metacognitive study habits, while study efficiency is driven mainly by GenAI usage frequency. The authors identify four learner profiles and conclude that the cognitive-skill/study-efficiency trade-off is not inevitable when intensive GenAI use is structured and metacognitively scaffolded.

Key findings

  • Best-performing models explained 28.18% of variance in Cognitive Skill (Gradient Boosting R² = 0.2818) and 30.00% of variance in Study Efficiency (Gradient Boosting R² = 0.3000).
  • SHAP analysis found Self-testing Habits (SHAP = 0.128) was the strongest predictor of cognitive skill development, followed by Study Notes via GenAI (SHAP = 0.120) and Active Learning (SHAP = 0.109); GenAI Usage Frequency (SHAP = 0.184) was the dominant driver of Study Efficiency.
  • OLS predicting critical thinking gain found Academic Writing via GenAI was the strongest predictor (β = 0.170, p < 0.001), followed by Creative Writing (β = 0.129) and Confidence/Motivation (β = 0.101); Anxiety/Worry and Positive Affect were non-significant.
  • K-means identified four profiles: Augmentors (high skill, high efficiency, highest GenAI use; n = 2,900), Optimisers (moderate balance; n = 5,500), Traditionalists (efficient with lower GenAI use; n = 3,000 in the abstract; ~3,200 in Table 8), and Substitutors (disengaged, low performance; n = 2,000). Institutional funding type was a significant predictor of study efficiency gains.

Methods

  • Secondary analysis of a publicly available cross-sectional multinational survey dataset: final analytical sample N = 14,949 higher education students, drawn from an original sample of 22,963 respondents across 120 countries (65.1% retained).
  • Five-layer analytical framework: (1) CFA and Fornell-Larcker validity testing; (2) multivariate OLS predicting critical thinking gain; (3) benchmarking of Linear Regression, Random Forest, XGBoost, and Gradient Boosting on a stratified 70/30 train-test split (training n = 9,548; test n = 4,092); (4) SHAP analysis on the best-performing model using a 500-participant test subset; and (5) K-means clustering (k = 4) across Cognitive Skill, Study Efficiency, and GenAI Usage.
  • Robustness checks used five-fold cross-validation on the full modelling dataset (n = 13,640) and discipline-stratified subgroup models for STEM (n = 1,301) and Social Science (n = 6,025).

Limitations

  • The paper states that cognitive skill development was measured through self-reported Likert items (Q29a–Q29e), a proxy measure; future research should use objective indicators such as GPA trajectories, standardised tests, or pre/post designs.
  • The paper states that the findings are predictive associations, not causal evidence, and cannot rule out selection effects; longitudinal panel designs are needed. It also notes geographic imbalance in the sample and that no formal comparison between included and excluded participants was performed, so selection bias cannot be completely excluded.
  • Evident from the text: the design is cross-sectional and used convenience sampling; cluster sizes are approximate and differ between the abstract and Table 8 for Traditionalists (n = 3,000 vs. ~3,200); the STEM subgroup is much smaller (n = 1,301) and shows lower predictive performance (R² = 0.144) than Social Science (R² = 0.241) or the full sample (R² = 0.282).

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s44163-026-02082-6

Cites

74 of 74 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Conceptual Framework. Source : Researchers' Construct (2026)
Fig. 1 Conceptual Framework. Source : Researchers' Construct (2026)From the PDF
Fig. 2 Cognitive skills vs. study efficiency trade-off analysis
Fig. 2 Cognitive skills vs. study efficiency trade-off analysisFrom the PDF
Fig. 3 Country-level comparison analysis. Note : Country-level mean scores for Cognitive Skill (navy blue), Study Efficiency (green), and GenAI Usage Intensity (amber) across the 10 largest national subsamples. Red dashed lines represent the cross-country grand mean. All scores on a 1-5 Likert scale.
Fig. 3 Country-level comparison analysis. Note : Country-level mean scores for Cognitive Skill (navy blue), Study Efficiency (green), and GenAI Usage Intensity (amber) across the 10 largest national subsamples. Red dashed lines represent the cross-country grand mean. All scores on a 1-5 Likert scale.From the PDF

Verification

Claims the paper makes

7 Valid · 1 Not valid · 44 Not checked · 89 Own results and interpretations

71 more claims · 1 with their cited sources checked

Create a free account to see all 141 claims and to verify the 1 reference still open.

See all claims — free account
  • Probably not a substantial claim. 42
  • It may state two separate claims as one. 35
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

5 of 5

Citations

Cites

Show all 65 references

Research works printed but not linked: 1

  1. Oyetade K, Zuva T. Advancing Equitable Education with Inclusive AI to Mitigate Bias and Enhance Teacher Literacy. Educational Process: Int J. 2025;14:e2025087. https://www.ceeol.com/search/article-detail?id=1338892.

Metadata

Authors
Bolatimi, O. S., Ricky-Okine, C., Bolatimi, F. E., Agbesi, C. C. M., Martey, F., Bolatimi, O. R., Peter, A., & Kwaku, E. O.
Affiliations
Koforidua Technical University (GH)
Venue
Discover Artificial Intelligence
Published
2026-08-28
Volume
6
Issue
1
ISSN
2731-0809 (ISSN-L)
OpenAlex
W7204577620
Cited by
0
References
65
FWCI
0.00
Type
journal article
Language
en
Open access
Licence
Locations
  • Discover Artificial Intelligence (publishedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Topics
  • Social Sciences › Education › Artificial Intelligence in Education
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Physical Sciences › Computer Science › Computer Science Applications › Online Learning and Analytics
Keywords
generative AI in higher education, cognitive enhancement, cognitive offloading, metacognition, learning efficiency, critical thinking, student interaction patterns, AI usage frequency, student engagement

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Multinational machine learning investigation of whether generative AI interactions augment or substitute cognitive skills and study efficiency in global higher education. https://app.cochat.ai/papers/10.1007/s44163-026-02082-6
The paper
Bolatimi, O. S., Ricky-Okine, C., Bolatimi, F. E., Agbesi, C. C. M., Martey, F., Bolatimi, O. R., Peter, A., & Kwaku, E. O. (2026). Multinational machine learning investigation of whether generative AI interactions augment or substitute cognitive skills and study efficiency in global higher education. Discover Artificial Intelligence. https://doi.org/10.1007/s44163-026-02082-6

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.