A draft framework to help organizations incorporate trustworthiness considerations into the design and implementation of generative artificial intelligence is a welcome start, but the document is too dense and could use further refinement, the American Bankers Association said in comments to the National Institute of Standards and Technology.
NIST is developing the voluntary framework as part of an executive order by President Biden directing federal agencies to implement policies to promote the safe deployment of AI technologies in the U.S. economy. In its comments, ABA said its members found the draft framework for generative AI very comprehensive, and that it adequately captures risks created or exacerbated by the technology. However, some of the information in the document is laid out “in an extremely dense manner,” the association said. ABA also suggested refinements to the risk types identified in the document so they better suit the specific needs and challenges encountered by financial institutions.
“These materials will serve as an ideal starting point for businesses looking to stand up their governance and will also be a benchmark for maturing existing practices,” ABA said.
