AI evidence extraction

ChatGPT in the Classroom: Student Perceptions and Learning Experiences in Higher Education

Nosipho Mavuso, Fezile Treasure Matsebula · International Journal of Learning Teaching and Educational Research · 2026

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

6/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.26803/ijlter.25.2.39

Methodology & findings

Study design

Mixed-methods sequential explanatory design combining survey data (n=248) with Focus Group Discussions (n=36).

Sample

N = 284, 2 groups

Primary method

Descriptive statistics combined with Technology Acceptance Model (TAM)-based thematic analysis. Specific statistical software and tests are not mentioned in the abstract.

Main result

The study found that "students generally view ChatGPT positively for increasing engagement, stimulating creativity, and supporting personalised learning, though perceived benefits vary across disciplines." Additionally, "students highlighted its usefulness for understanding complex subjects and completing assignments, but raised concerns about academic integrity, overreliance on Artificial Intelligence (AI), and potential erosion of critical thinking."

Reports effect sizes.

Research paradigm

Mixed-methods (pragmatism)

Author conclusions

The study emphasises that "while ChatGPT can enrich learning, its integration requires deliberate pedagogical planning. In South African higher education, educators must establish guidelines for the ethical use of AI, embed AI literacy into curricula, and design assessments that promote authentic learning." The results suggest the need for "balancing innovation with academic integrity and fostering dialogue between students and educators to responsibly harness the potential of ChatGPT."

Risk of bias

Convenience sampling (non-probability) introduces selection bias; Limited to two South African universities may limit generalizability; Purposive sampling for FGDs may introduce selection bias; Self-reported perceptions subject to social desirability bias; Selection bias: Convenience sampling for survey (248 participants) reduces representativeness; Selection bias: Purposive sampling for FGDs (36 students) may not represent broader student population; Potential volunteer bias: Self-selected participation in focus groups; Limited geographic scope: Only two South African universities, limiting generalizability; No reported attrition rate or dropout analysis; No reported inter-rater reliability for thematic analysis; Convenience sampling for survey (248 participants) introduces selection bias; Purposive sampling for FGDs (36 students) may not be representative; Non-probability sampling methodology limits generalizability; Study limited to two South African universities, restricting geographic and institutional generalizability

Open questions raised

The authors identify the need for: (1) establishment of guidelines for ethical use of AI in higher education; (2) embedding AI literacy into curricula; (3) designing assessments that promote authentic learning; (4) fostering dialogue between students and educators regarding responsible AI use.
The study identifies the need for: (1) establishment of guidelines for ethical use of AI in higher education; (2) integration of AI literacy into curricula; (3) design of assessments promoting authentic learning; (4) ongoing dialogue between students and educators regarding responsible ChatGPT use
Future research should examine implementation of AI literacy in curricula, develop assessment strategies that promote authentic learning while leveraging ChatGPT, and investigate disciplinary differences in ChatGPT integration across different academic fields.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 76%

Explore related topics

Responsible AI Use Human-AI Collaboration Mixed Methods