Operationalizing AI ethics in medicine-a co-creation workshop study.

Area of Science:

• Artificial Intelligence Ethics
• Human-Computer Interaction
• Participatory Design

Background:

• Existing AI ethics frameworks often lack practical, actionable guidance for implementation in real-world projects.
• Operationalizing AI ethics requires translating abstract principles into concrete project requirements.
• Co-creation workshops offer a novel approach to bridge this gap.

Purpose of the Study:

• To propose and evaluate a novel method for operationalizing AI ethics through co-creation workshops.
• To elicit low-level ethical requirements from stakeholders for an AI-based clinical decision support system.
• To identify ethical requirements not covered by existing EU guidelines.

Main Methods:

• A qualitative, participatory research approach involving diverse stakeholder teams.
• Workshops utilized storytelling, content analysis, and Planguage for requirement elicitation.
• Procedural feedback mechanisms were employed for iterative refinement.

Main Results:

• Key ethical issues identified include explainability, privacy, robustness, validity, fairness, and transparency.
• Concrete, quantifiable requirements were drafted for privacy, explainability, transparency, and validity.
• Six critical issues, including time sensitivity and patient harm prevention, were not mapped to EU guidelines.

Conclusions:

• The co-creation workshop methodology effectively elicits actionable, project-specific AI ethics requirements.
• The method is resource-efficient and adaptable, requiring fewer AI ethics experts.
• Further training in Planguage and addressing power dynamics are recommended for optimal outcomes.