AI evidence extraction

Unveiling Students’ Experiences and Perceptions of Artificial Intelligence Usage in Higher Education

Xue Zhou, Joanne Zhang, Ching Chan · Journal of University Teaching and Learning Practice · 2024

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

7/10

Relevance

3/4

Quality (LMQS)

Evidence

Citations

29.96

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.53761/xzjprb23

Methodology & findings

Study design

Mixed methods approach combining online surveys and semi-structured interviews.

Sample

N = 28, 6 groups

Primary method

Descriptive statistical analysis on closed-ended survey responses. Thematic analysis using Braun and Clarke's (2006) six-phase framework: familiarisation with data, initial coding, theme searching, theme reviewing, theme defining and naming, and producing report. Inductive coding. Independent theme review by all authors for cross-checking and rigor.

Main result

The study found that students employ AI tools for diverse purposes including writing assistance (80% using Grammarly), text summarization (20%), administrative tasks (40%), market research (7%), and advertising (10%). Key benefits identified include "increased productivity, personalized learning, and enhanced linguistic capability." However, "concerns regarding academic integrity, over-reliance on AI, and the need for clear usage guidelines are also identified." Students reported that approximately one-third improved their writing skills since using AI writing-assistant tools, and 50% responded positively to AI's capacity to serve as an intelligent tutor.

Reports effect sizes.

Research paradigm

Mixed methods (positivist and interpretivist)

Author conclusions

"In summary, this exploratory study combined surveys and in-depth interviews to provide empirical insights into how students utilise AI in their self-directed learning, including perceived benefits and challenges. Findings help address gaps in understanding actual versus speculative student practices. Results may inform institutional policies and pedagogies on effectively integrating AI tools in HE." Furthermore, the authors conclude that "clear guidance and training on how to use AI in HE is urgently required."

Risk of bias

Selection bias: Purposive sampling based on AI use may not capture students with no AI experience equally; Small sample size (n=28) from single institution limits generalizability; Attrition potential: 12 of 16 surveyed students participated in interviews; Potential recall bias in interview responses; Non-random sampling of entrepreneurship students only; Selection bias: purposive sampling based on AI use status; Sample size bias: only 28 participants from single institution; Self-selection bias: voluntary participation may skew toward students interested in AI; Social desirability bias: students may overstate benefits or understate concerns in interviews; Selection bias: purposeful sampling based on AI use; potential self-selection bias of volunteers; single institution limits generalizability; small sample size (n=28); no control group; researcher positionality not explicitly stated.

Limitations

"The primary limitation of this exploratory study was its restricted sample size from a single institution." The study was conducted at one university (Queen Mary University of London) with 28 total participants, limiting generalizability
Sample size of 28 is appropriate for exploratory qualitative research but insufficient for broader conclusions.

Open questions raised

Lack of comprehensive studies examining how AI technologies support Self-Directed Learning (SDL) principles in entrepreneurship education
Limited empirical research on whether generative AI is adopted by students and how they use it in entrepreneurship education
Need for larger-scale research with wider sample sizes encompassing entrepreneurship students from various institutions and levels
Gap between speculative benefits and evidence-based studies on AI effectiveness in supporting learning needs
The authors identified that "there is a lack of comprehensive studies examining how AI technologies specifically support SDL principles to enhance personalised learning in educational settings, particularly in the context of entrepreneurship education." They also note that "empirical research remains limited on whether generative AI is being adopted by students, and if so, how students use AI to assist their learning in entrepreneurship education." Future research should employ "a wider sample size, encompassing entrepreneurship students from various institutions and levels."
Gap in AIEd literature: lack of comprehensive empirical studies examining how AI technologies specifically support self-directed learning (SDL) principles in entrepreneurship education. Authors note most discussion consists of speculative benefits rather than evidence-based studies. Call for larger-scale research: "Future research could employ a wider sample size, encompassing entrepreneurship students from various institutions and levels."

Data: Not mentioned as available; No datasets explicitly mentioned as available for sharing.Code: None mentioned; None mentioned.Extracted from: pdfAgreement 46%

Explore related topics

Responsible AI Use Human-AI Collaboration Mixed Methods