AI evidence extraction

Pedagogy with generative artificial intelligence: Opportunities and challenges in education

Ojo Amos Adewale, Jayesh Rane, Martina Oluchi Ogbonna, Nitin Liladhar Rane · International Journal of Applied Resilience and Sustainability · 2026

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

7/10

Relevance

0/4

Quality (LMQS)

Evidence

Citations

0.00

FWCI

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.70593/deepsci.0202013

Methodology & findings

Study design

The paper presents a mixed-methods empirical study comparing student performance with and without ChatGPT across AI-assisted assignments and proctored examinations, combined with student perception surveys measuring AI dependence and learning concerns..

Sample

4 groups

Primary method

Inferential statistics (t-tests or similar, based on p-values reported); descriptive statistics for student perceptions (percentages: 71%, 58%)

Main result

The study found that "Students using ChatGPT demonstrated significantly superior performance on AI-assisted assignments (p < 0.001)" with benefits including time efficiency and personalized learning experiences. However, "the AI group scored lower on proctored examinations (p < 0.01), suggesting potential overreliance that may hinder independent learning capabilities." Additionally, "71% acknowledging AI dependence for task completion while 58% expressed concerns about compromised learning depth."

Reports effect sizes and confidence intervals.

Research paradigm

Mixed methods (quantitative and qualitative)

Author conclusions

The authors conclude that generative AI functions as a "double-edged sword" and advocate that "generative AI's transformative potential, including personalized tutoring, immediate feedback, and content generation efficiency can only be realized through deliberate pedagogical innovation" requiring "comprehensive educational reform encompassing redesigned assessments that prioritize genuine understanding over AI-generated responses, institutional policy frameworks governing ethical AI use, and systematic AI literacy development among students and educators."

Risk of bias

Selection bias: students self-selected into AI-assisted group vs control; Assessment method bias: take-home assignments may favor AI use; Hawthorne effect: awareness of study may influence behavior; Confounding variables: prior academic ability, motivation, technology comfort not explicitly controlled; Selection bias: unclear how students were assigned to AI vs. non-AI groups; Confounding: different assessment types (take-home vs. proctored) may not be directly comparable; Measurement bias: student self-reported perceptions may be subject to social desirability bias; Attrition: no information provided on dropout rates or missing data; Selection bias: students self-selected into AI use groups; Measurement bias: different assessment types (proctored vs. take-home) may favor different learning modalities; Confounding: student motivation, prior knowledge, and technology comfort not explicitly controlled; Temporal bias: cross-sectional design limits causal inference

Limitations

The abstract indicates that "Traditional assessment methods, particularly take-home assignments, proved inadequate in AI-enabled environments, highlighting the need for fundamental pedagogical restructuring," and the study notes the need for "longitudinal studies examining sustained learning outcomes, discipline-specific AI applications, and development of evidence-based guidelines" to address gaps in current understanding.

Open questions raised

Longitudinal studies examining sustained learning outcomes
Discipline-specific AI applications
Development of evidence-based guidelines balancing educational enhancement with academic integrity preservation
Comprehensive educational reform encompassing redesigned assessments and institutional policy frameworks
Systematic AI literacy development among students and educators
The authors call for "longitudinal studies examining sustained learning outcomes, discipline-specific AI applications, and development of evidence-based guidelines balancing educational enhancement with academic integrity preservation."

Data: not_statedCode: not_statedExtracted from: pdfAgreement 59%

Explore related topics

Productivity Effects Doctoral Training Experimental