NTIA recommends independent audits, regulatory inspections for high-risk AI systems

The federal government should establish accountability for emerging artificial intelligence systems at all stages of development by increasing transparency and requiring independent audits or regulatory inspections for those presenting a high level of risk, according to a long-awaited report from the National Telecommunications and Information Administration.

“The report calls for improved transparency into AI systems, independent evaluations of those systems, and consequences for imposing new risks,” NTIA Administrator Alan Davidson told reporters on a March 26 call ahead of its...