AI evidence extraction

Responsible Utilisation of AI in Academic Research Strategies for Preserving Scientific Integrity and Quality

Iman Osman Mukhtar Ahmed, Farida Tadjine, Aisha Hassan Abdalla Hashim · Journal of Engineering Research and Education (JERE) · 2026

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

9/10

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/10.58915/jere.v18.2026.3021

Methodology & findings

Study design

Qualitative exploratory approach using multi-document analysis as the primary data collection method.

Main result

The study found that "AI systems rely on data-driven algorithms; however, they struggle to comprehend context when encountering novel material or situations that fall outside their training parameters" and that "AI is incapable of drawing conclusions or making decisions in complex circumstances." Additionally, the research revealed that "the use of artificial intelligence in scientific writing lacks essential skills such as critical thinking, originality, creativity, and ethics, which could negatively impact students' research writing abilities," and that "many academics are using AI and its tools without adequately reviewing the data obtained, which can result in plagiarism" including "data falsification, plagiarism, data fabrication, and excessive reliance on AI-generated (GTP) chat tools."

Research paradigm

Interpretivist/qualitative exploratory

Author conclusions

"An intelligent researcher is one who knows how to utilise AI tools in a balanced manner, understanding the concept, developing it in his own style, and adding to it with his own knowledge." The authors conclude that "Using these tools is acceptable, but it's illogical for a researcher to be a mere transmitter, not thinking about or reviewing what is presented to him" and emphasize that "This balance guarantees that the final result adheres to the utmost standards of academic research while benefiting from the speed and precision offered by AI tools."

Risk of bias

Selection bias in document analysis (peer-reviewed sources may not represent all AI usage contexts); Language bias in AI systems trained predominantly in English; Potential confirmation bias in thematic analysis coding; Limited representation of non-English language AI tool effectiveness; Reliance on secondary sources rather than primary empirical data collection; Selection bias: Multi-document analysis may not represent all perspectives on AI in research; source selection criteria not explicitly stated; Confirmation bias: Use of AI tools (ChatGPT, ChatPDF) to analyze literature about AI may introduce bias regarding AI capabilities; Publication bias: Reliance on peer-reviewed sources may exclude critical perspectives on AI in research; No explicit discussion of researcher positioning or potential conflicts regarding AI adoption advocacy; Selection bias in document sampling (peer-reviewed sources only, may exclude gray literature); Language bias (focus on English-language AI tools and English-language research); Potential confirmation bias in thematic analysis; Limited representation of non-English-speaking research contexts

Open questions raised

The authors identify needs for: (1) development of AI systems specifically designed for academic research rather than commercial multifunctional tools; (2) improved effectiveness of AI in non-English contexts and languages such as Arabic; (3) better methods to detect AI-generated content and distinguish human-written from AI-generated texts; (4) verification methodologies for AI-sourced references; (5) international standards for AI ethics and regulation; (6) enhanced training programs for researchers on AI prompt engineering and ethics; (7) discipline-specific criteria for AI use in different research fields; and (8) methodologies to mitigate bias in AI training datasets.
Need for discipline-specific AI guidelines tailored to unique requirements of each scientific field
Necessity to establish robust norms promoting ethical AI development with diverse data ensuring fairness and inclusivity
Requirement for methodologies to detect AI-generated content and differentiate between human and AI-produced texts
Development of international standards for AI ethics and regulation to control fabrication, falsification, misinformation, and bias
Need for collaborative frameworks between countries and scientific institutions to develop coherent international ethical standards for AI research

Extracted from: pdfAgreement 74%

Explore related topics

Responsible AI Use Research Integrity Qualitative