New Bill Would Codify Limits on Government AI Use
U.S. Sen. Elissa Slotkin (D-Mich.) introduced legislation that would ban the federal government from using artificial intelligence for mass surveillance of Americans and restrict AI's role in military decisions. The AI Guardrails Act of 2026 comes after a public dispute between AI company Anthropic and the Department of Defense over how the government deploys AI technology.
The bill prohibits three specific uses: AI-aided mass surveillance targeting American individuals or groups, AI systems making lethal force decisions without human authorization, and AI independently launching or detonating nuclear weapons.
What triggered the legislation
Anthropic, which provides AI technology to government agencies, raised concerns about how the Pentagon was using its systems, particularly for domestic surveillance and autonomous weapons. The Department of Defense said it already had guidelines prohibiting these uses.
Anthropic remained skeptical that existing guidelines contained adequate safeguards. In response, the Pentagon designated Anthropic a supply chain risk and instructed all federal agencies to terminate contracts with the company. Anthropic is now suing over that designation.
The dispute has cost taxpayers money. Agencies must now remove Anthropic software from their systems, and the legal battle could extend the expenses further.
Why codification matters
Slotkin, a member of the Armed Services Committee, argued that if these restrictions were written into law rather than left as internal guidelines, the conflict might have been prevented. Guidelines lack the force of statute and can shift with administrations.
The bill will also prompt broader discussions about AI contracts in the National Defense Authorization Act, the annual defense spending bill scheduled for later this year.
For government employees overseeing technology procurement or policy, understanding these regulatory developments is essential. Learn more about AI for Government and how policy frameworks are evolving, or explore the AI Learning Path for Policy Makers to deepen your knowledge of governance and oversight in this space.
Your membership also unlocks:
