Senate Republican seeks to fast-track AI whistleblower bill

Senator Chuck Grassley, the Iowa Republican who has spent decades building a reputation as Congress’s most persistent whistleblower advocate, introduced the AI Whistleblower Protection Act on May 15, 2025. The bill, designated S. 1792, would create federal protections for AI industry employees who flag security vulnerabilities or safety concerns to regulators, shielding them from the kind of corporate retaliation that currently keeps many workers quiet.

The legislation arrives at a moment when AI companies are racing to deploy increasingly powerful systems while simultaneously locking down internal dissent with aggressive non-disclosure agreements and restrictive severance packages. Grassley’s bill takes direct aim at those contractual muzzles, seeking to ensure that the people closest to potentially dangerous AI systems can actually speak up without torpedoing their careers.

What the bill actually does

At its core, S. 1792 establishes anti-retaliation protections for workers who report significant AI-related security risks. The bill has attracted bipartisan co-sponsors, including Senator Chris Coons, a Delaware Democrat, and Senator Marsha Blackburn, a Tennessee Republican.

The central problem the legislation attempts to solve is straightforward. AI companies routinely require employees to sign NDAs that can extend well beyond trade secret protection into territory that effectively prevents workers from reporting safety issues. Severance agreements often come with similar strings attached. Grassley’s bill would make it illegal to retaliate against employees who choose to report concerns. It would also void NDA provisions that attempt to prevent protected disclosures, creating a legal safe harbor for workers who bring concerns to designated federal agencies.

The legislative traffic jam

Despite its bipartisan backing, S. 1792 has been referred to the Senate Committee on Health, Education, Labor, and Pensions, where it has stalled with no further action recorded as of September 2026.

The bill’s stalling contrasts with the trajectory of a related piece of legislation, the Stop Stealing Our Chips Act, which focuses on whistleblower protections specifically in the context of semiconductor export controls. That bill managed to pass in May 2026, suggesting that Congress is more comfortable with narrowly scoped whistleblower measures tied to national security concerns than with broader protections covering the AI industry writ large.

Supporters of the Grassley bill, including the National Whistleblower Center, continue to push for its advancement.

Why the AI industry’s NDA culture matters

Some separation agreements have included provisions that could claw back vested equity if former employees make disparaging statements, a category broad enough to potentially encompass legitimate safety disclosures. Without a clear federal standard establishing that safety-related disclosures are protected, employees are left navigating a legal gray zone where the safest career move is silence.

The fragmentation of AI policy efforts in Congress, with semiconductor export controls moving forward while broader safety measures stall, points to a deeper challenge. Lawmakers appear willing to act on AI issues when they can be framed as national security imperatives, but slower to move when the primary beneficiaries are individual workers rather than geopolitical strategy.