WASHINGTON, D.C.—Federal agencies have pitched more than 1,000 use cases for generative artificial intelligence, but the U.S. government’s cybersecurity chief wants risks associated with the technology properly addressed before they can press ahead.

Generative AI tools such as OpenAI’s ChatGPT and Google’s Bard have piqued the interest of companies and government agencies alike, for their ability to parse large data sets and answer conversational queries. But as with any technology on government radar, caution is in order, said Chris DeRusha, federal chief information security officer at the Office of Management and Budget, speaking at a FAIR Institute conference of cybersecurity and risk managers.