EPIC Submits Additional Feedback on NIST AI Risk Management Framework 

EPIC has submitted feedback to NIST to inform the development of an AI Risk Management Framework that will assist companies in developing, deploying, and evaluating trustworthy AI systems. This is the second time that EPIC has submitted comments to NIST on this issue and comes in response to an updated draft of the Framework. EPIC’s feedback recognizes the important progress that NIST has made to expand the types of AI risks and harms covered by its Risk Management Framework, including risks to racial equity and to those impacted by automated decision-making, and to include more diverse voices in the AI development process. However, EPIC urges NIST to (1) prioritize meaningful accountability mechanisms that keep companies honest, (2) incorporate additional protections for those impacted by high-risk AI practices, including data minimization standards, protections for sensitive user data, and protections against unsubstantiated AI practices like emotion regulation and biometric categorization, and (3) provide clear examples of recommended practices that help companies comply with the Framework. 

EPIC frequently advocates for algorithmic justice, transparency, and accountability, and has recently provided comments on the D.C. Council’s Stop Discrimination of Algorithms Act, the European Commission’s proposed Artificial Intelligence Act, and the OECD Framework for Classifying AI Systems.