Risk, Rule-setters and Rule-takers: Regulatory approaches to risk in AI-supported and AI-automated decision-making for general welfare
Focus Areas: News and Media, Transport and Mobility, Health, and Social Services
Research Programs: Institutions
This project seeks to scope several approaches to deal with Automated Decision-Making and Decision-Support Systems-Related Risks (ADM/DSS RR) through norms and provide an evaluation of those approaches for their consideration in regulatory contexts.
The standpoint is to look at risk control of those systems beyond ethics or social principles and focus the discussion on the possible interventions from the regulator’s perspective.
The overarching questions and sub-questions guiding this project are:
- What is risk in an ADM / DS System?
– Is it possible to define it?
– How is it different from technological risk?
– How is it different from the concept of risk in
– Are all “potential harms” risks of and ADM/DSS?
– Is there a concept of risk usable for regulatory
- What types of risks are common and which ones specific to ADM/DSS?
– Due to the nature of the risk?
– Due to the scale of the risk?
- What is an acceptable risk:
– From the point of view of society as a whole
– From the point of view of the most vulnerable groups
– From the point of view of the owner of the AI system
– From the point of view of the users of the system
- Can risk be separated from questions of liability/ responsibility or do they need to be regulated together?