12,445 papers · continuously updated · last export: 10 Aug 2026livingmeta.ai
← Browse all papers
AI evidence extraction

Tracing 40 years of research on Artificial Intelligence and human metacognition from 1985 to 2024

Mahima Anna Varghese, Poonam Sharma · Discover Psychology · 2025

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

7/10
Relevance
0/4
Quality (LMQS)
C
Evidence
1
Citations
2.38
FWCI
Top 10%
Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1007/s44202-025-00463-z

Methodology & findings

Study design

Bibliometric analysis using VOSviewer and Bibliometrix (R-package) software.

Sample

N = 144, 1 group

Primary method

Bibliometric tools including VOSviewer and Bibliometrix (R-package) were used for performance analysis, co-authorship analysis, and co-word analysis of publication data.

Main result

The analysis reveals "a gradual increase in publications with AI and human metacognition as common point of study from 1985, with a sharp rise post-2009 and a peak in 2023." The United States leads in research output, and "the most cited work is by Graesser (2005), focusing on metacognitive scaffolding through intelligent tutoring systems such as iSTART and AutoTutor."

Reports effect sizes.

Research paradigm

Positivist/Empiricist (bibliometric analysis of scientific literature)

Author conclusions

The authors conclude that "Theoretically, this research repositions AI not only as a facilitator of technology but as a cognitive collaborator that directly influences metacognitive activities such as planning, monitoring, and reflective judgment. This theoretical framework helps to develop metacognitive theory in the digital era by situating AI as a collaborator in, not a substitute for, human cognitive control."

Risk of bias

Selection bias: Single database (Scopus) may miss grey literature and non-indexed publications; Language bias: Likely English-language bias in indexed publications; Publication bias: Positive results more likely to be published and indexed; Time-dependent bias: Earlier periods (1985-2000) may have lower publication rates due to historical factors rather than research interest

Open questions raised

  • The study identifies a gap regarding empirical research on negative effects of AI on human metacognition, noting limited discussion of overreliance on AI and fear of failure. Future research should examine how AI intermediates with human metacognition beyond facilitator roles.
  • The study identifies the need for more empirical research on negative effects of AI on human metacognitive abilities, including overreliance on AI and fear of failure.
  • The study identified limited empirical research on the negative effects of AI on human metacognitive abilities, including topics such as overreliance on AI and fear of failure. The authors suggest that future research should examine AI as a cognitive collaborator in the digital era rather than solely as a technology facilitator.
Data: not_statedCode: not_statedExtracted from: pdfAgreement 70%

Explore related topics

Productivity EffectsHuman-AI CollaborationMeta-Analysis

Related papers

← Browse all papers