As the Pentagon rushes to integrate artificial intelligence into everything from intelligence analysis to battlefield operations, lawmakers are calling for limits on the Pentagon’s use of artificial intelligence.
A new version of the annual defense policy bill could include AI guardrails, including a ban on the Pentagon launching autonomous systems to kill people without their permission, spy on Americans or launch nuclear weapons.
Democratic Sen. Elissa Slotkin told NOTUS on Tuesday that she expects to introduce the bill. Introduced in March It will be incorporated into the basic text of the National Defense Authorization Act of 2027. The Senate Armed Services Committee is expected to weigh in on the bill next week.
“The base text is [bipartisan] “If it had been controversial, this deal wouldn’t have been here. So I took this as a real sign that people knew we needed left-right limits, limits that were reasonable and not overly authoritarian,” the Michigan senator said.
Sen. Kirsten Gillibrand on Tuesday introduced a bill along similar lines that would prohibit the Pentagon from using AI to launch nuclear weapons, conduct domestic surveillance of Americans, or deploy most fully autonomous weapons.
“Right now, the Department of Defense is moving toward deploying incredibly powerful AI technologies without common-sense guardrails in place, with potentially devastating consequences that will make us all safer,” Gillibrand said in a statement about her bill, the Safe and Accountable Military AI Act. “We must act now. Rather than stifle technological progress, we must put humans in charge and establish clear rules to keep the use of AI in warfare wise and safe.”
President Donald Trump is pushing for rapid adoption of AI across the federal government while backing away from some proposed safety measures. president trump tuesday signed an executive order It directs various federal agencies to strengthen the cybersecurity of critical infrastructure. It requires AI companies to voluntarily submit new models to a 30-day government review before making them available to the public.
On the other hand, Secretary of Defense Pete Hegseth argues that: national security at stake The military is pushing to rapidly deploy and integrate artificial intelligence. confidential network. the effort It spurred a fight with Anthropic, one of the developers of the technology.after taking a moral position that supports ethical limits.
“We judge AI models by these criteria alone: factually accurate, mission-relevant, and free of ideological constraints that limit legitimate military use,” Hegseth said at a SpaceX event in January. “The War Department’s AI will never wake up. It works for us. We’re building weapons and systems for war, not chatbots for Ivy League faculty lounges.”
However, there are many skeptics of AI both inside and outside the military.
Pope Leo XIV I called recently The government has called for strict regulation of AI, declaring that it is “unacceptable” to rely on AI to make fatal decisions.
Vice President J.D. Vance took a similar line, using remarks at the Air Force Academy last week to urge service members not to cede decision-making to AI.
“In some ways AI will positively change the battlefield, but in some ways it won’t, so I want people to be jealous and selfish about their role as decision-makers in war.” Mr. Vance said in his commencement address:.
Some military leaders are wary of how artificial intelligence will be used. At a recent press conference, Gen. Frank Bradley, commander of U.S. Special Operations Command, said: The US military recently said “We need to be very careful” about the use of AI.
AI may be able to identify potential targets, he said, but “we humans need to have confidence that it will only bring violence where it is intended.”
Gillibrand’s bill would create new limits on how the Pentagon uses artificial intelligence for “high-consequence” applications such as nuclear, lethal targeting, domestic surveillance, and cyber, and would require Pentagon leadership to first approve and notify Congress.
It also establishes department policy that AI supports, but does not replace, human judgment in decisions involving enforcement, detention, home surveillance, or other high-impact AI applications. The bill generally bans fully autonomous weapons, but allows semi-autonomous systems for missile defense and target tracking, identification and prioritization.
