AI-based research mentors: Plausible scenarios and ethical issues

Area of Science:

• Ethics in Artificial Intelligence
• Technology Foresight
• Research Mentorship

Background:

• AI-based research mentors (AIRMs) are emerging tools in scientific research.
• Ethical implications of AIRMs require thorough examination.
• Anticipatory Technology Ethics provides a framework for foresight and issue evaluation.

Purpose of the Study:

• To examine AI-based research mentors (AIRMs) using technology foresight.
• To identify and evaluate ethical issues associated with different AIRM scenarios.
• To inform policy development for responsible AIRM implementation.

Main Methods:

• Scenario planning was used to develop four plausible future scenarios for AIRM use.
• Resnik's twelve principles of research ethics were applied to identify ethical issues.
• Ethical issues were weighed against potential benefits in each scenario.

Main Results:

• Openness, education, legality, and mutual respect were violated in all AIRM scenarios.
• Freedom was compromised in scenarios involving AIRM monitoring.
• The "AIRMing" phenomenon demonstrated potential manipulation for poor practice in the guidance scenario.

Conclusions:

• The guidance-only AIRM scenario presents the fewest ethical problems.
• Even the guidance-only scenario raises concerns regarding honesty, openness, credit, education, legality, and mutual respect.
• Policy is essential to ensure AIRMs are used solely for guidance and to mitigate identified ethical issues.