AI evidence extraction

Generative Ai and Student Engagement, Motivation, and Learning: Exploring The Use Of Generative Ai in Higher Education Online Courses

Jeffrey L. Simmons · 2026

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

6/10

Relevance

0/4

Quality (LMQS)

Evidence

Citations

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.22371/05.2026.002

Methodology & findings

Study design

Mixed-methods study combining quantitative Likert-scale surveys with qualitative thematic analysis.

Sample

< 30, 2 groups

Primary method

One-sample t-tests against a neutral midpoint, cross-course comparisons using several different inferential techniques, Spearman rank-order correlations, and thematic analysis of open-ended responses.

Main result

The study found that "Results indicated high overall levels of student engagement (M=4.31) and perceived learning (M=4.28). However, motivation varied significantly depending on the AI tool's pedagogical positioning. Courses that used NotebookLM as a 'study companion' yielded significantly higher intrinsic motivation than courses that used ChatGPT as a transactional 'productivity tool.'" Additionally, "AI-generated audio features (podcasts) effectively managed cognitive load, supporting Mayer's Cognitive Theory of Multimedia Learning."

Reports effect sizes and confidence intervals.

Research paradigm

Mixed methods (positivist/interpretivist)

Author conclusions

The authors conclude that "institutions should frame AI as a collaborative partner to foster intrinsic motivation and enable personalized learning at scale rather than solely as an efficiency engine." They emphasize that "The study concludes that GenAI serves as an effective cognitive scaffold, enabling students to prevent cognitive overload while maximizing germane load processing and functioning as the 'more capable peer' within students' Zone of Proximal Development."

Risk of bias

Selection bias: participants from single institution; Sampling bias: voluntary post-course survey completion; Instruction bias: different AI tools and pedagogical positioning across courses; Potential social desirability bias in survey responses; Selection bias: Single institution study (small private nonprofit university in California); Limited sample scope: Only three online master's-level courses examined; Potential social desirability bias: Post-course surveys may be subject to student response bias; Lack of comparison/control condition: No courses without AI integration for baseline comparison; Instructor prediction bias: Faculty expectations may influence course design in ways that affect outcomes; Single institution study - selection bias; Self-report survey data - response bias; No control group - unable to establish causality; Pre-course instructor predictions - expectancy effects; Voluntary survey completion - attrition and non-response bias

Limitations

The authors explicitly state that "Although the findings have limited generalizability, they may help address gaps in course design and development in this new frontier of AI integration in universities." The study was conducted at a single small private nonprofit university in California across three online master's-level courses, which constrains the broader applicability of findings.

Open questions raised

The authors identify the need to address gaps in course design and development for AI integration in universities and call for research to shape course-level AI integration policies and guidelines for effective use of large language models in online higher education.
The authors identify that findings address "gaps in course design and development in this new frontier of AI integration in universities" and note the need for broader research to shape "course-level AI integration policies and guidelines for the effective use of large language models in online higher education." The limited generalizability indicates a need for larger, multi-institutional studies.
Need for broader generalizability beyond single institution
Further investigation of how AI pedagogical positioning affects motivation across diverse institutional contexts
Exploration of long-term learning outcomes beyond perceived learning measures
Assessment of how different student populations respond to varied AI integration approaches

Data: not_statedCode: not_statedExtracted from: pdfAgreement 62%

Explore related topics

Responsible AI Use Human-AI Collaboration Mixed Methods