Comments on NIST's GenAI profile

The National Institute of Standards and Technology (NIST) recently released draft documents relating to their work on President Biden’s Executive Order on AI.

Among them was the AI RMF Generative AI Profile (NIST AI 600-1). This guidance document helps developers of GenAI systems implement the NIST’s AI Risk Management Framework.

The GenAI profile provides a great overview of 12 of the key risks to society from GenAI, as well as hundreds of actions that organizations can start implementing to better manage societal AI risk.

My colleagues and I submitted our comments on the profile here. We suggest additional risks to add to the risk list, ways to categorize the risks and best practices for sound risk taxonomies. We also provide suggestions of ways to categorize and add risk management actions.