Washington, D.C – A coalition of veterans and consumer advocacy groups, including the Consumer Federation of America and Mozilla, issued a letter urging the White House to maintain critical safeguards on AI systems that impact millions of Americans. A new Executive Order could allow federal agencies to use untested, unregulated AI to decide who gets VA healthcare, who clears airport security, and who receives retirement benefits, without requiring basic transparency or testing.

“Transparency and testing to make sure AI systems funded by taxpayers work without bias or mistake is a non-partisan no-brainer,” says Ben Winters, Director of AI & Privacy at the Consumer Federation of America. “We’re hopeful the administration will retain these commonsense oversight mechanisms that protect both people and federal agency investments.”

“Mozilla has been advocating for trustworthy AI approaches for years, recognizing that AI can be harmful when it’s not built and deployed safely, said Nik Marda, Technical Lead for AI Governance at Mozilla. “Not only has there been a longstanding bipartisan consensus that AI should be tested before it’s used, but that’s also best practice from a technical perspective too. We hope that the White House’s review of this policy will keep in place these most basic and essential protections for federal use of AI.”

Rollbacks of these requirements would cause real harm to veterans, seniors, and consumers while wasting taxpayer dollars on tech that simply doesn’t work. The coalition is demanding the administration maintain those key provisions when revising regulations before it’s too late.