No autonomous system should make a decision to kill without a human being who answers for it. Due process and accountability depend on a person being responsible for the use of force. When that responsibility shifts to an algorithm, no one can be held accountable when it's wrong.

There is currently no federal law requiring this. In 2026, a federal agency retaliated against a contractor for refusing to remove exactly these safeguards, a designation a federal court later found unconstitutional.

As Senator, Lydia will:

• Require human authorization for any lethal force decision made using AI in any DoD contract, with no exceptions for classification or urgency
• Prohibit federal agencies from using "supply chain risk" or similar national security designations to retaliate against a contractor for refusing to remove constitutional safeguards from a contract
• Require public disclosure whenever a federal agency asks a contractor to weaken human oversight or privacy protections as a condition of a contract