Table 2.

Summary of “Call for Action”-items for regulators and organization derived from the workshop discussions.

Provide clear definition of risks	Risks arising from and with the application of AI are manifold and multidimensional. A clear and standardized definition of ‘risks’ is inevitable to allow for effective risk governance.
Provide standardization regarding risk governance	Currently a large variety of risk governance approaches exists. Unification and standardization are important to give guidance on which risks to address and how.
Provide clear accountability frameworks	A lack of clear definition of responsibilities and accountabilities for risks and their management was expressed.
Generate transparent, widely understandable and practical methodologies	Several downsides were seen across currently proposed risk management methods. Particularly their transparent and practical application along the full process chain was regarded challenging.
Include human impact evaluations in risk management processes	Evaluation regarding impacts of risks for humans was expressed to be only partially included or insufficient. Increasing the importance of human impact analyzes throughout risk assessment and management techniques has been demanded.
Provide clarity about how to handle unforeseen events	Due to their unpredictability unforeseen events and risks arising with them are still challenging to solve in practice. Clarity regarding responsibility and how to deal with them is required.