US Lawmakers Float Federal AI Safety Board to Oversee High‑Risk Models

In Washington, a bipartisan group of legislators is reportedly drafting a bill to create a new federal body dedicated to AI safety, tentatively framed as an “AI Safety Board.” The proposed agency would focus on high‑risk systems such as large language models and autonomous AI agents, overseeing how they are developed, deployed and monitored.

According to early descriptions, the board would function as an independent regulator, loosely modeled on the National Transportation Safety Board. It would investigate serious AI incidents, define benchmarks and evaluation protocols for powerful models, and help set procurement standards for federal agencies. The intent is to complement, rather than replace, the work already being done by the FTC and NIST.

Industry groups are already voicing concern that a standalone AI regulator could slow innovation and add overlapping red tape. With the US presidential election only months away, it is unclear how far the bill will advance this year. Still, the fact that Congress is openly debating a permanent AI watchdog marks a shift from the previous, guidance‑heavy approach to AI policy at the federal level.