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

Learning to Learn with Gen AI: A Qualitative longitudinalstudy

Sehrish Javed, Mehmood Chadhar, Anitra Goriss-Hunter · Journal of the Association for Information Systems · 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)
E
Evidence
0
Citations
0.00
FWCI

Methodology & findings

Study design

Longitudinal qualitative study using semi-structured interviews conducted over one year, analyzed through thematic analysis approach.

Primary method

Thematic analysis approach for qualitative data analysis. No quantitative statistical methods reported.

Main result

The study identified three key themes in how students use Gen AI: "(1) an output-oriented study approach; (2) Rethinking the rationale behind the outcome of Gen AI tools; (3) focusing on using Gen AI for their own understanding and learning." The research demonstrates that "students transitioned from an over-reliance on Gen AI tools to the conscious use of them to support their learning, i.e., for cognitive augmentation rather than replacement" and that "Gen AI tools became tools for collaboration rather than tools for substitution."

Reports effect sizes.

Research paradigm

Interpretivist/Qualitative

Author conclusions

The authors conclude that "This study highlights the importance of structured training and ethical guidelines to ensure Gen AI tools are used as facilitators" and that the "research contributes in understanding the transformative potential of Gen AI in education while addressing its challenges. These challenges include over-reliance and blind trust. Findings provide actionable insights for educators and policymakers to optimize Gen AI integration, ensuring it supports long-term student development and equitable learning opportunities."

Risk of bias

Selection bias (participants may be self-selected based on interest in Gen AI); Attrition risk over one-year longitudinal period; Potential interviewer bias in semi-structured interview collection; Potential social desirability bias in self-reported learning experiences; Selection bias: Participants were self-selected higher education students from diverse disciplines; Attrition risk: Longitudinal study over one year may have participant dropout; Social desirability bias: Semi-structured interviews may elicit socially desirable responses about Gen AI use; Selection bias: Participant self-selection into longitudinal study may introduce motivation bias; Attrition risk: Longitudinal design over one year carries risk of participant dropout; Interviewer bias: Semi-structured interviews subject to interviewer effects and interpretation bias; Recall bias: Participants retrospectively reporting on learning experiences; Limited generalizability: Qualitative sample from higher education may not represent broader student populations

Open questions raised

  • The authors identify that "the long-term implications of its use on student learning, effectiveness and efficiency remain underexplored" prior to this study.
  • The study addresses the gap that "While its short-term benefits are well-documented, the long-term implications of its use on student learning, effectiveness and efficiency remain underexplored." Future directions include addressing challenges of "over-reliance and blind trust" in Gen AI tools.
  • The authors identify that "while its short-term benefits are well-documented, the long-term implications of its use on student learning, effectiveness and efficiency remain underexplored." They also note challenges including "over-reliance and blind trust" that need to be addressed.
Data: not_statedCode: not_statedExtracted from: pdfAgreement 67%

Explore related topics

AI Literacy TrainingDoctoral TrainingQualitative

Related papers

← Browse all papers