As artificial intelligence pushes deeper into military systems, a US lawmaker is moving to set firm limits on how it can be used in combat.
US Senator Elissa Slotkin has introduced legislation aimed at placing legal boundaries on military AI, including bans on autonomous lethal decisions, AI-driven surveillance of Americans, and use in nuclear weapons launches, according to NBC News.
The proposal would codify existing Pentagon guidelines on AI use, giving current safeguards legal force.
“Our bill is a neat five pages. This is not an extensive, elaborate thing,” Slotkin told NBC News.
The move follows recent tensions between the Pentagon and Anthropic, which has resisted loosening guardrails on its AI models over concerns they could be used for surveillance or lethal targeting.
Rising Scrutiny Over Military AI
The push for tighter limits on military AI reflects broader public unease about the growing role of AI in warfare.
A YouGov–Economist poll found that 44 percent of Americans support limits on military AI use, while only 25 percent believe the tech should be deployed without restrictions.
Similar concerns also emerged within the tech industry, where employees have called for safeguards against mass surveillance and autonomous weapons.
“We understand that, like with every tool ever invented, there are some really good uses that help and there are some really dangerous uses,” Slotkin said.
The legislation is expected to feed into early debates over the next National Defense Authorization Act, Congress’s annual defense policy bill.
