Agentic education · Learning science

Actionable companion · Assessment & reasoning

Learning with agency: Amplifier or substitute? A systematic review of generative AI’s impact on higher-order cognitive skills among university students

AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not claims or recommendations attributed to the source authors. No independent peer review of this guide is asserted.

Read original paperSearch paper agentCC BY 4.0

Evidence and context

Title and abstract screening; stage labels describe reported settings, not verified numerical ages. Design labels require full-text appraisal. Reported stage(s): Higher education. Verify the participant population and task before transferring the findings.

Original authors’ abstract, reproduced under the linked source license:

Background As generative artificial intelligence (GenAI) becomes embedded in university learning environments, understanding its effects on students’ higher-order cognitive development has become one of the most pressing questions in educational research. Methods This systematic review was conducted in accordance with PRISMA 2020 guidelines, synthesizing 89 peer-reviewed journal articles published between 2024 and 2026 from Web of Science and Scopus. Narrative synthesis was employed to integrate findings across studies. Results GenAI’s cognitive effects were neither uniform nor unconditional: positive outcomes were documented in 40.4% of studies and mixed or conditional effects in 23.6%. ChatGPT was the predominant tool examined (n = 61, 68.5%), and mixed-methods designs were most prevalent (42.7%). Critically, 55.1% of studies employed no specified pedagogical strategy, and explicit theoretical frameworks were identified in only 25.8% of the corpus. Over-reliance emerged as the leading cognitive risk (33.7%), followed by reduced analytical autonomy (20.2%) and cognitive offloading (18.0%). Discussion These findings support a Dual-Mechanism Model: GenAI functions as a cognitive amplifier under structured pedagogical conditions and as a cognitive substitute under unguided use. The evidence calls on universities to anchor GenAI integration within deliberate instructional frameworks that preserve students’ cognitive agency and foreground autonomous higher-order reasoning.

What kind of evidence is this?

Supporting AI learning science · Evidence synthesis

Stage: Higher education. Screening labels describe reported context and study-design language. They are not a risk-of-bias rating or proof of effectiveness for an age band.

Study methods · source passage

This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 guidelines (Page et al., 2021), which provide the highest standard for transparent, reproducible, and comprehensive evidence synthesis. The systematic review design was selected as the most appropriate framework for synthesizing the heterogeneous and rapidly expanding body of empirical evidence on GenAI’s effects on university students’ critical thinking, creative thinking, and problem-solving skills. To promote methodological transparency, all eligibility criteria, search strategies, and synthesis procedures were specified and documented prior to data collection, minimizing the risk of post-hoc reporting bias. This review was not prospectively registered in PROSPERO or any equivalent international register prior to commencement. This is acknowledged as a methodological limitation and is addressed further in Section 5.

Limitations and uncertainty · source passage

This review is subject to several limitations that should be considered when interpreting its findings. As a single-researcher design, structured mitigation protocols were implemented—including dual-pass screening, time-delayed re-appraisal of a 15% random subsample, and two-occasion data extraction—yet these cannot fully replicate the inter-rater reliability achievable in multi-reviewer designs. The review was not registered in PROSPERO prior to data collection; given the rapidly evolving nature of the GenAI education landscape and the intentionally narrow temporal scope of 2024–2026, prospective registration was not pursued, as the protocol was fully documented, and the review timeline did not permit pre-registration without compromising the currency of the evidence base.

Design the system before the pilot

Use this study to inform a learning mechanism or implementation question. It is not direct evidence that autonomous agents improve learning. Keep the proposed agent behavior separate from what the source actually tested.

  • Use common rubrics, blind scoring where feasible and a fresh unaided reasoning task. Document permissible assistance and avoid treating AI-detector scores as proof of misconduct.

A 90-day district pilot

  1. Days 1–15 · Appraise. Have the curriculum lead and a teacher read the methods and limitations. Record the study design, sample, comparator, outcomes and evidence gaps. Define a local question that the study can help inform.
  2. Days 16–30 · Design. Choose one disciplinary task. Require a first attempt, a documented AI consultation, a critique of the output and an independent explanation. Align the comparison group and scoring rubric with the same objective. Name an accountable owner and set the baseline, success threshold and budget before implementation.
  3. Days 31–60 · Test. Start with a small voluntary cohort, preserve an accessible non-AI route and require teacher review. Use synthetic or de-identified examples; approve data handling and tool terms through district processes.
  4. Days 61–90 · Decide. Blind-score reasoning and accuracy on a new task without AI. Distinguish model examination scores, student perceptions and actual learner gains; only the last directly tests educational benefit. Compare with baseline, document uncertainty and stop or revise if learning worsens, access is unequal or critical errors persist.

The board decision brief

Report the learning problem, evidence quality, participating population, total cost including review time, measured benefit, errors, access differences and remaining uncertainties. Make a dated recommendation to stop, revise or expand. A positive perception survey alone does not establish learning improvement.

Attribution and reuse

Fawzia Omer Alubthane. Amplifier or substitute? A systematic review of generative AI’s impact on higher-order cognitive skills among university students. Frontiers in Psychology (2026-06-22). https://doi.org/10.3389/fpsyg.2026.1863931

Companion prepared 2026-10-30. Original authorship and DOI are preserved. The abstract is an attributed source extract; the implementation plan is new AI-assisted editorial material, offered under the same CC BY 4.0 license. This guide has not undergone independent peer review.