Sen. Adam Schiff (D-Calif.) introduced legislation Monday that would require human involvement in the Pentagon's use of autonomous or semi-autonomous weapons systems powered by artificial intelligence. The bill, called the Human Authority in Lethal Operations Act, seeks to establish a clear chain of command under which a designated commander retains final authority over decisions to use force involving AI-driven weapon systems.

The proposal would also mandate that Pentagon officials maintain detailed records of decision-making processes such as target selection for post-action review. Additionally, the legislation includes provisions barring military use of AI to monitor individuals engaging in constitutionally protected activities, removing human involvement from nuclear weapons deployment, and restricting purchases of personal data on Americans from third parties.

"The past few months have shown us that there is an urgent need for commonsense guardrails to ensure the Defense Department's use of AI is in line with Americans' national security and privacy priorities," Schiff said in a statement. "There are good reasons to use AI technology to advance our national security, however — just as with any tool, we cannot depend on technology alone to guide us, particularly when the risks of harm can be fatal."

Schiff's bill arrives as several other Senate Democrats pursue similar constraints on military AI through amendments to the National Defense Authorization Act. Sen. Kirsten Gillibrand (D-N.Y.) introduced her own legislation last week that would prohibit the Pentagon from using AI to launch nuclear weapons, surveil Americans, and deploy autonomous weapons without human oversight. She has indicated plans to offer her measure as an NDAA amendment.

Sen. Elissa Slotkin (D-Mich.) also reportedly intends to introduce separate AI guardrails legislation as an amendment to the annual defense policy bill. The coordinated Democratic push follows a highly publicized dispute between the Pentagon and AI company Anthropic earlier this year over terms governing use of the company's AI models for military applications.

What the Left Is Saying

Progressive Democrats framing the issue emphasize ethical imperatives and constitutional safeguards. Schiff's office pointed to concerns about AI systems making irreversible life-and-death decisions without human judgment. Civil liberties advocates aligned with the Democratic caucus have long argued that autonomous weapons raise profound moral questions about accountability when a machine, rather than a commander, makes a killing decision.

Supporters of the legislation note that current Pentagon policy already requires "appropriate" human involvement in lethal autonomous systems under a 2023 directive, but argue that language is insufficiently specific. Schiff's bill would codify those requirements into law with clearer enforcement mechanisms and record-keeping mandates.

Democratic sponsors have also highlighted privacy provisions prohibiting AI surveillance of constitutional activities as essential to preventing mission creep. The data-purchase restrictions target what civil libertarians describe as a growing Pentagon practice of acquiring commercially gathered information on Americans without warrants.

What the Right Is Saying

Conservative defense hawks have expressed concern that prescriptive AI restrictions could place unnecessary bureaucratic friction between commanders and their forces during combat operations. Some Republican national security analysts argue that adversaries like China and Russia face no such constraints, potentially giving them operational advantages in future conflicts.

A spokesperson for Senate Armed Services Committee Republicans noted that existing Pentagon directives already address human oversight requirements while preserving operational flexibility. The committee has historically resisted legislative micromanagement of military technology deployment, preferring to leave such determinations to defense professionals.

Other Republican voices have taken a more nuanced position, supporting human-on-the-loop requirements for nuclear systems and certain autonomous weapons while opposing blanket restrictions they characterize as overly broad. These lawmakers argue that the definition of "autonomous" versus "semi-autonomous" systems remains legally ambiguous, potentially capturing defensive technologies like missile interceptors already in widespread use.

What the Numbers Show

The Pentagon budget for AI-related programs exceeded $1.7 billion in fiscal year 2025, according to Defense Department spending reports. That figure has roughly tripled since 2020 as the military accelerates integration of machine learning systems into intelligence analysis, logistics planning, and weapons guidance.

A 2024 Government Accountability Office audit found that 17 major weapon systems under development incorporated some form of autonomous target recognition capability. Of those, six had no formal requirement for human confirmation before engagement — a gap the GAO recommended addressing through updated acquisition regulations.

Public polling from the Brennan Center for Justice indicates 67 percent of Americans oppose allowing AI to make lethal decisions without human review, while 23 percent support autonomous weapons under specific battlefield conditions. Support for human oversight rises to 84 percent when nuclear systems are included in survey questions.

The Bottom Line

The Schiff bill represents the most comprehensive legislative attempt yet to impose statutory guardrails on military AI use. Its fate likely depends on whether sponsors can attract Republican co-sponsors willing to accept operational restrictions in exchange for bipartisan credibility heading into fall NDAA negotiations.

Three factors will determine the legislation's trajectory: the outcome of ongoing Pentagon-Anthropic negotiations that prompted renewed scrutiny, whether Gillibrand and Slotkin amendments gain traction in committee markup, and how the Congressional Budget Office scores any record-keeping mandates. Defense contractors with existing autonomous weapons contracts are monitoring developments closely for potential contract modification requirements.

Congress returns from recess February 24, when NDAA drafting committees will begin formal consideration of AI-related amendments. Schiff's office indicated the bill text would be formally introduced and referred to the Senate Armed Services Committee upon that return.