Senator Elissa Slotkin (D-MI), a member of the Senate Armed Services Committee, introduced the AI Guardrails Act on Tuesday, March 17, 2026 — a concise five-page bill that would impose strict statutory limits on the Department of Defense’s use of artificial intelligence in military operations.
Wisdom Imbibe Insight:
The AI Guardrails Act signals a critical turning point: governments are no longer just racing to build powerful AI—they’re beginning to define its boundaries. In warfare, the question is no longer what AI can do, but what it should never be allowed to do. The future of military power may depend as much on restraint as innovation.
The legislation seeks to:
- Prohibit autonomous lethal strikes without meaningful human authorization and oversight.
- Ban AI-enabled mass surveillance of U.S. persons and citizens on domestic soil.
- Explicitly forbid the use of AI to launch, detonate, or authorize nuclear weapons.
Slotkin described the bill as a necessary first step to establish clear “left and right limits” on military AI before the technology outpaces congressional oversight.
“Congress is behind in putting left and right limits on the use of AI, and the first place to start should be at the Pentagon,” Slotkin said in a press release. “AI is going to shape the future of America’s national security, and we must win the AI race against China. But to do that, we need action that puts limits on AI in the Department of Defense.”
The bill would codify two existing DoD policy guidelines — human-in-the-loop requirements for lethal force and prohibitions on domestic mass surveillance — while adding the new nuclear weapons restriction as a statutory red line.
Context: The Anthropic-Pentagon Clash as Catalyst
The timing is no coincidence. The bill arrives amid the high-profile dispute between the Pentagon and AI company Anthropic. After Anthropic refused to remove contractual restrictions on using its Claude models for mass domestic surveillance or fully autonomous weapons, Defense Secretary Pete Hegseth designated the company a “supply chain risk” in March 2026. President Trump then ordered federal agencies to cease using Anthropic technology, prompting the company to file lawsuits challenging the designation.
Slotkin explicitly tied her legislation to this fallout, arguing that clear statutory guardrails could have prevented the entire episode:
“The Pentagon targeted Anthropic in this situation and will spend the next year, along with an unknown number of taxpayer dollars, removing Anthropic from all classified systems — a dispute that could have been resolved with proper legislation,” she told NBC News.
The ongoing U.S. military campaign against Iran (Operation Epic Fury) has further heightened scrutiny, with AI tools — including Palantir platforms that previously incorporated Claude — assisting in target identification and operational planning. Slotkin emphasized the need for “human redundancy” in high-stakes decisions, expressing skepticism about current assurances.
Broader Congressional Momentum
Slotkin introduced the bill as a standalone measure but aims to fold its provisions into the 2026 National Defense Authorization Act (NDAA), the annual must-pass defense policy and spending bill. The legislation currently has no co-sponsors, but parallel efforts are underway:
- Senator Adam Schiff (D-CA) is drafting separate legislation focused on autonomous weapons and domestic surveillance.
- Senator Mark Kelly (D-AZ) is collaborating with colleagues on AI governance language for the NDAA.
These efforts reflect growing bipartisan concern that the Pentagon’s rapid AI adoption — especially in targeting, intelligence analysis, and decision support — lacks sufficient statutory boundaries, even as the U.S. races to maintain technological superiority over China.
Outlook and Challenges
The AI Guardrails Act is unlikely to pass as standalone legislation in the current Congress but could influence NDAA negotiations, where defense hawks and civil liberties advocates often clash. Supporters argue it provides clarity and prevents future vendor disputes; critics (including some defense officials) may view it as overly restrictive or premature given the classified nature of many AI applications.
As the Anthropic lawsuit proceeds and AI’s role in active conflicts like Iran continues to expand, Slotkin’s bill marks the first formal congressional attempt to draw hard red lines around military AI — signaling that lawmakers intend to play a more active role in shaping how the Pentagon wields this transformative technology. Whether it gains traction will depend heavily on the NDAA process and the evolving political landscape around national security and civil liberties in 2026.
Recommended for you: