Anticipating risks and promoting quality in AI for health: A participatory workshop 

The PD-TIPS.AI project aims to co-design a bilingual conversational recommendation system that provides people living with Parkinson’s disease with personalized self-care advice and resources. Phase 1 mapped out the needs and user profiles, as well as the initial design principles for the system.

Looking ahead to Phase 2, the project must anticipate the potential ethical, social, clinical and organizational risks associated with the use of AI and develop a quality control and risk management framework.

The objective of the proposed project is to design and facilitate a participatory workshop to:

• collectively identify the risks associated with a conversational recommendation system in health;
• examine the system’s quality control, validation and governance mechanisms; and
• produce a key deliverable (a risk and quality control framework) to serve as the basis for Phase 2 of the project and for future simulations. 

The project is grounded in a participatory, sociotechnical approach that combines:

• a targeted review of the literature on the risks of AI in health, responsible AI and the governance of algorithmic systems;
• co-design and critical foresight methods (e.g. risk mapping);
• the design of a structured workshop (objectives, activities and materials); and
• a qualitative analysis of workshop discussions to identify operational principles and recommendations.

Students will be involved in preparing the methodology and structuring the analysis of the results. 

This project will enable students to develop:

• a critical understanding of the ethical and social issues raised by AI in health;
• risk-analysis skills;
• skills in facilitating and designing participatory workshops;
• the ability to translate collective discussions into analytical frameworks; an
• skills in writing strategic and scientific deliverables. 

Activity Number of hours

• Targeted literature review (responsible AI, risks, quality): 20 hours
• Analysis of materials from Phase 1 of the PD-TIPS.AI project: 15 hours
• Preparation, organization and facilitation of the participatory workshop: 15 hours
• Development of tools (materials, matrices, scenarios): 15 hours
• Analysis of the workshop results and development of the framework: 15 hours
• Meetings with the research team: 10 hours

Total: 90 hours