Democrats in Congress are pushing to pass bills that would limit the Department of Defense’s use of artificial intelligence.

This week, Democratic Sen. Adam Schiff from California unveiled the Human Authority in Lethal Operations (HALO) Act, a bill that would require a human commander to have the final say on any course of action taken by autonomous weapon systems. The bill would also mandate detailed record-taking of how military decisions were made and targets were selected for later review, establish protections for whistleblowers, and prohibit the use of AI in some cases involving nuclear weapons and mass surveillance.

“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,” Senator Schiff said in the press release. “My legislation would protect Americans from unlawful domestic surveillance, ensure that humans in the chain of command exercise responsibility for the use of any lethal technology, and maintain strong ethical protections in the deployment of autonomous and semi-autonomous weapons.”

Artificial intelligence has been a part of warfare for some time now. Militaries around the world employ the help of AI systems for target selection in strikes and mass surveillance, one prominent example being the use of AI by the Israeli army against Palestinians. The United States has also long deployed artificial intelligence in military operations, including in its latest war against Iran.

But earlier this year, the use of AI in the military was catapulted to the top of public discourse when an existing deal between the Pentagon and Anthropic fell through, and in an unprecedented move, the AI giant was designated a supply chain risk. Anthropic had allegedly refused to get rid of guardrails in its AI systems that were meant to prevent the DoD from using its models for mass domestic surveillance and completely autonomous weapons, meaning little to no human involvement.

Following the fallout of the deal, the Pentagon signed contracts with pretty much all the other major AI companies, including OpenAI, Google, Nvidia, SpaceX, Microsoft, and Amazon Web Services. Meanwhile, Anthropic has challenged the designation in court, though reports say the Trump administration is warming up to the company following the release of Mythos, its latest model advertised as a nightmare for the cybersecurity industry.

After the Trump administration’s very public breakup with Anthropic, a slew of Democrats came out in support of the AI company and its stance. That list included Sen. Schiff, the author of the HALO Act.

“I wish we had more voices like Anthropic out there,” Sen. Schiff said at the Punchbowl News Conference in March.

Schiff has introduced a range of AI-related bills in the past few months, including proposals to require large data centers to pay for their own power, mandate AI companies to disclose copyrighted work used to train models, and bring AI literacy classes to schools. Now, he is reportedly aiming to tag this bill along to the annual military spending package, aka the NDAA, which must pass by the end of the year at the latest.

He is not the only Democrat with a plan like this. Sen. Kirsten Gillibrand of New York introduced a very similar bill earlier this month, bringing restrictions to the use of AI in nuclear weapon deployment, domestic surveillance and fully autonomous weapons. All of these and more “high-consequence actions,” as the bill defines them, would need approval from a high-ranking DoD official to move forward. Sen. Gillibrand is also reportedly planning on introducing the proposal as an amendment to the NDAA.

Then, there’s the AI Guardrails Act introduced by Sen. Elissa Slotkin of Michigan back in March. Also aiming for very similar guardrails to Sen. Schiff and Gillibrand’s bills, Sen. Slotkin is reportedly preparing to introduce it as an amendment to the NDAA as well.

While all three of these bills aim to ensure safety by bringing human oversight to any decision made by AI systems in military settings, that’s not where the dangers end.

Many AI users tend to suffer from what experts call automation bias, aka believing that an AI system can make more accurate judgments than you because it has access to more information or perhaps reasons in a more efficient way. That, obviously, is not true: the technology is far from perfect, and LLMs are prone to hallucinations or biased thinking. Combine that with the “black box” nature of AI, where users don’t have complete insight into how or why the system reasons the way it does, and you have a military AI plan that could still lead to potentially fatal mistakes even with human oversight.