AI evidence extraction

Towards Ethical AI Adoption in Academic Research: Insights from a Systematic Literature Review

Faheem Ullah, Babar Shah, Ahmed Saeed Alshehhi, Abrar Ullah, Sajid Anwar · Proceedings of the AAAI Symposium Series · 2026

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

9/10

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/10.1609/aaaiss.v9i1.42913

Methodology & findings

Study design

Systematic literature review following PRISMA framework.

Sample

N = 16, 3 groups

Primary method

Thematic synthesis rather than statistical meta-analysis was employed due to heterogeneity in methodologies of included studies. Boolean operators (AND, OR, NOT) and field-specific syntax were used in database searches. CASP checklist was adapted for quality appraisal.

Main result

The review found that "existing studies consistently identify bias, transparency, accountability, and data privacy as the primary ethical risks associated with AI-assisted research" and that "while AI tools improve research speed and analytical capacity, they also raise ethical concerns that affect research quality, academic integrity, and fairness." The analysis reveals "persistent concerns about algorithmic bias, transparency deficits, and accountability gaps, which vary in manifestation across disciplines but collectively threaten research integrity."

Reports effect sizes.

Research paradigm

Mixed-methods interpretive review with thematic synthesis

Author conclusions

The authors conclude: "This systematic review has synthesized the current discourse on ethical considerations in AI-assisted academic research, addressing three core dimensions: scientific research practices, higher education contexts, and emerging ethics tools. The analysis reveals that while AI offers transformative potential for research efficiency and innovation, it simultaneously introduces complex ethical challenges that demand interdisciplinary solutions." They further state: "The academic community must foster ethical AI literacy while developing adaptive accountability mechanisms to maintain scholarly rigor in this rapidly changing landscape."

Risk of bias

Database selection bias with overrepresentation of STEM fields; Publication bias favoring positive outcomes regarding AI in research; Underrepresentation of social sciences and humanities perspectives; Methodological heterogeneity across included studies limiting comparability; Limited geographic and institutional diversity in included studies; Potential inclusion bias toward English-language publications only; Risk of missing emerging ethical challenges due to technology evolution; Absence of articles prior to 2023 limits historical perspective

Limitations

The authors state: "Although multiple databases were consulted, the literature is dominated by studies from STEM fields
Our database selection and search strategy naturally prioritized repositories such as IEEE Xplore, ACM Digital Library, and arXiv, where AI research is most actively published
As a result, perspectives from the social sciences and humanities may be underrepresented." Additional limitations include: "The largely qualitative nature of the reviewed work also limits generalization across institutions and regions." The authors also note "rapid changes in AI technologies mean that some ethical challenges discussed in earlier studies may already be evolving."

Open questions raised

Existing studies focus on technical applications and overlook ethical issues specific to academic research practices
Limited agreement on how principles such as fairness, accountability, and transparency should be applied within AI-supported research workflows
Role of higher education institutions in developing ethical AI literacy among researchers and students is insufficiently examined
Effectiveness and suitability of AI ethics tools for scholarly settings have not been rigorously evaluated
Few studies take an interdisciplinary view mapping issues across different domains
Need for longitudinal studies of ethical challenges in evolving AI systems

Extracted from: pdfAgreement 64%

Explore related topics

Responsible AI Use AI Governance Literature Review