AI evidence extraction

Generative AI in Higher Education: Uses and Ethical Dilemmas of Students

Meital Amzalag, Gila Kurtz · Artificial intelligence · 2025

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

9/10

Relevance

3/4

Quality (LMQS)

Evidence

Citations

2.91

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.5772/intechopen.1012802

Methodology & findings

Study design

Mixed-methods study combining quantitative survey data (online questionnaire with Likert scales) and qualitative content analysis of open-ended responses.

Sample

N = 673, 14 groups

Primary method

Quantitative analysis: SPSS software for frequencies, averages, standard deviations, independent samples t-tests, and Mann-Whitney U test (non-parametric alternative to t-test). Qualitative analysis: content analysis conducted in five stages with inductive theme identification, inter-researcher reliability assessment (initial 60%, improved to 63%, then 83% after refinement of themes).

Main result

The study found that "GenAI tools, particularly ChatGPT, have become integrated into students' academic routines, with undergraduate students using these tools more frequently than graduate students." Students reported using GenAI to search for learning materials, clarify complex content, and generate inspiration. Undergraduate students use GenAI tools for educational purposes more frequently (M = 2.869, SD = 0.926) than graduate students (M = 2.438, SD = 0.945), with a statistically significant difference (t(428) = 3.309, p = 0.001). The findings reveal that "students expressed concerns regarding the accuracy of information, privacy, academic integrity, and the absence of clear institutional guidelines."

Reports effect sizes and confidence intervals.

Research paradigm

Mixed methods (positivist quantitative + interpretivist qualitative)

Author conclusions

The authors conclude: "This study makes a significant contribution by providing empirical insights into the extent and purposes of GenAI tools for learning among students in higher education institutions in Israel, as well as the ethical dilemmas they encounter. The findings highlight the use of GenAI tools, such as ChatGPT, in students' academic routines. Moreover, the findings underscore the importance of providing targeted support and guidance, particularly for undergraduate students who rely more heavily on GenAI tools for learning purposes." They further state: "Our work highlights the need for higher education institutions to adapt their curricula to the new realities of the GenAI era. This includes providing students with the knowledge, skills, and ethical frameworks necessary to navigate and critically engage with GenAI technologies."

Risk of bias

Selection bias: voluntary nature of participant responses led to overrepresentation of women (72.2% vs. population average) and underrepresentation of graduate students (9.5% vs. population proportion); Language bias: English-based GenAI tools used by non-native English speakers (Israeli participants), potentially affecting usage frequency and patterns; Response bias: self-reported usage data without objective verification; Convenience sampling in pilot phase (n=289); Selection bias: Sample bias in favor of women (72.2% vs. expected population proportion) and against graduate students (9.5% vs. expected proportion) due to voluntary response nature; Language bias: All GenAI tools tested were English-based, not native language of Israeli participants; Sampling method: Convenience sampling in pilot phase, though stratified sampling used in final national study; Measurement bias: Inter-researcher reliability initially low (60%), improved to 83% after refinement of coding themes; Recall bias: Self-reported GenAI usage without behavioral verification; Selection bias: Voluntary participation led to overrepresentation of women (72.2%) and undergraduates (86.3%); Language bias: English-based GenAI tools used by non-native English speakers may have affected usage frequency and patterns; Sampling limitation: Convenience sampling in pilot study (n=289); Reporting bias: Reliance on self-report questionnaire data without behavioral verification; Attrition/completeness: Open-ended question responses (n=116) represented only 24.6% of GenAI users (471); respondents were self-selected; Qualitative analysis reliability: Initial inter-rater reliability was only 60%, improved to 63% after revision, eventually reaching 83% after theme refinement

Limitations

The authors state: "In this chapter, due to a minor bias in the representation of women and respondents with a bachelor's degree, we recommend interpreting the study's findings with caution and suggest replicating this manuscript with a sample that more accurately reflects the population's demographics
Furthermore, we cannot determine whether the use of GenAI tools, primarily for completing assignments such as writing articles, solving math exercises, and writing papers, constitutes negative uses by students or if these actions were based on instructions received from their lecturers
Additionally, since the qualitative data analysis was based on written responses to open-ended questions in the online questionnaire, we were unable to request clarifications or examples when responses were unclear." Further limitations noted include: "the reliance on a nonnative language interface (English-based GenAI tools) may have influenced both the frequency and nature of student usage."

Open questions raised

Limited knowledge about students' uses and ethical dilemmas when integrating GenAI into their learning across diverse contexts
Variation in students' ethical reasoning across different academic levels, prior experiences with GenAI, and diverse cultural and linguistic contexts remains understudied
Need for empirical studies exploring ethical implications of AI in educational settings
Need for semi-structured, in-depth interviews for deeper understanding of ethical reasoning
Need for studies after GenAI tools are introduced in native languages
Need for comparative research with student populations in different countries and institutional contexts with varying GenAI policies

Extracted from: pdfAgreement 56%

Explore related topics

AI Manuscript Writing Academic Writing Mixed Methods