Table 1.

Summary of points mentioned by participants during the discussion on pressing challenges regarding AI and accountability.

Acceptance and Trust	Deployment and technology acceptance are two different things. Trust is key in acceptance; thus, we need to demonstrate trust.
Data Bias	AI systems must not be biased against certain groups in society. Non-discrimination and data quality needs to be ensured during development and deployment.
Education	Education is key. People need to be educated on the risks and safety of AI, data scientists and developers need to be educated on the ethical challenges of AI, and regulators need to be educated on current technological developments.
Explainability	There is a gap between what can be explained and what needs to be explained. Additionally, it needs to be ensured that people can understand what the system explains.
Implications	Accountability needs to be understood in terms of how but also which systems to design. Only because we can do something does not mean we should do it.
Privacy	How can high data privacy standards be fulfilled in AI systems?
Regulation	Detailed legal acts and legal cases are required.
Safety and risk	Technology can never be 100% safe. The question is, how much risk is bearable, what is safe enough and how can we determine suitable thresholds.