AI guidelines should not create misalignment with existing regulations

Credit unions must have confidence that AI innovation will not be penalized by heavy-handed regulation, America’s Credit Unions’ Andrew Morris wrote to the National Institute of Standards and Technology (NIST) Friday. The comments were sent in response to NIST’s request for information on the White House’s October 2023 Executive Order on “Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.”

“While new standards and guidelines for AI can help achieve regulatory clarity, they should be tailored to avoid excessive burden, competitive imbalance between regulated and unregulated institutions, or the imposition of unreasonable supervisory expectations for credit unions,” Morris wrote , adding that NIST should “consult with federal banking regulators to determine how any future statements or best practices published in response to the E.O. might be interpreted in the context of existing financial regulatory frameworks.”

Morris added that the generality of certain sections of the executive order creates a high likelihood of overlap with existing laws and standards applicable to financial institutions, including the CFPB’s unfair, deceptive, and abusive acts or practices prohibition.

Development of standards for auditing and maintaining AI services, including chatbots, is “best reserved to the NCUA and other federal banking agencies whose supervisory authority puts them in a better position to understand their regulated institutions,” Morris argued.

Morris also encouraged NIST to consider developing standards to enable financial institutions to better understand actions taken by their regulators which are the result of “AI-driven decisions or assessments.”

America’s Credit Unions supports efforts to promote and encourage responsible innovation across all technological domains, including through the deployment of AI.

heelo