AI evidence extraction

Ethical Dimensions of Artificial Intelligence in Scientific Research: a Case Study of «7m01504 – Chemistry» Master's Program

Bibigul Dosanova, Adina Akhmetova, Zhadyra Akhmetova · InterConf · 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/10.51582/interconf.19-20.05.2026.002

Methodology & findings

Study design

Survey-based case study combined with qualitative interviews.

Sample

< 30, 5 groups

Primary method

Descriptive statistics (percentages) presented in Table 1. No inferential statistics, significance testing, or hypothesis tests reported.

Main result

The study found that "the majority of chemistry students (75%) use AI primarily as a supportive tool for technical tasks." However, a significant challenge emerged with the "False Positive" phenomenon, where students reported that "I wrote every sentence of my thesis myself, based on my laboratory experiments, but the system flagged it as 100% AI-generated," demonstrating that AI detectors often misidentify formally structured academic scientific writing as machine-generated content.

Reports effect sizes.

Research paradigm

Mixed methods (qualitative interviews + quantitative survey)

Author conclusions

The authors conclude that "The integration of AI into academic research is an irreversible process. However, to preserve the sanctity of the '7M01504 - Chemistry' program and the university's global reputation, the focus must shift from algorithmic policing to ethical empowerment. By adopting a transparent disclosure culture and acknowledging the limitations of AI-detectors, the academic community can harness technology to enhance human intellect."

Risk of bias

Selection bias: Study limited to one master's program (7M01504 – Chemistry) at one institution; Potential response bias in student interviews regarding AI usage patterns; Limited geographic scope: K. Zhubanov Aktobe Regional University (Kazakhstan); No explicit discussion of interviewer bias or interview protocol standardization; Selection bias: Only surveyed first-year master's students from one program (7M01504 - Chemistry); Potential sampling bias: No information on response rate or representativeness; Interviewer bias: Qualitative interviews without stated inter-rater reliability; Confirmation bias: Focus on a single institution may not represent broader practices; Selection bias: Sample limited to first-year master's students in one chemistry program at one institution; Self-reported data: Reliance on student survey responses regarding AI usage patterns; Potential social desirability bias in student interviews regarding academic integrity practices; No control group or comparison institution for contextual validation

Open questions raised

The paper identifies the need for: (1) a comprehensive regulatory framework for AI use in research; (2) implementation of disclosure protocols at institutional level; (3) shifting from algorithmic detection to substantive evaluation through oral defense; (4) development of Human-in-the-Loop approaches in scientific disciplines like Chemistry; and (5) better training of AI-detection systems to reduce false positives in formal academic writing.
The paper identifies the need for better implementation of disclosure protocols, shifting evaluation focus to substantive assessment through oral defenses and viva voce examinations, and adoption of Human-in-the-Loop (HITL) approaches where "Every AI-generated graph or formula must be manually verified against laboratory journals to ensure that AI hallucinations do not infiltrate scientific databases."
The paper identifies the need for better mechanisms to distinguish legitimate academic writing from AI-generated content, and calls for development of more sophisticated approaches beyond algorithmic detection, including implementation of disclosure protocols and human-in-the-loop verification systems.

Extracted from: pdfAgreement 63%

Explore related topics

Responsible AI Use Research Integrity Survey