URGENT: Oppose Pentagon Pressure to Bypass AI Privacy and Safety Guardrails

I am writing to express my grave concern regarding Defense Secretary Pete Hegseth’s current ultimatum to the AI company Anthropic. The Department of Defense is reportedly pressuring Anthropic to remove two critical guardrails by this Friday, February 27. I am asking you to intervene against these demands, which carry equal and devastating risks to our nation: 1. The Threat to Civil Liberties: Domestic Surveillance The Pentagon is demanding the removal of restrictions that prevent AI from being used for mass domestic surveillance. Utilizing advanced AI to monitor American citizens or scan private communications is a direct assault on the Fourth Amendment. We cannot allow national security tools to be turned inward against the American public. 2. The Threat to Tactical Safety: Autonomous Lethal Force Simultaneously, the Pentagon is demanding the removal of "human-in-the-loop" requirements for lethal force. Stripping these safety protocols from AI models—which are still prone to "hallucinations" and unpredictable errors—is a recipe for catastrophic failure, friendly fire, and unintended escalation on the battlefield. It is deeply disturbing that the Pentagon is threatening to label a pro-safety American company a "supply chain risk"—a designation usually reserved for foreign adversaries—simply because they refuse to enable domestic spying or unsafe weaponry. I urge you to: -- Convene Oversight Hearings: Investigate the weaponization of "supply chain risk" designations to bully American tech companies into dropping ethical standards. -- Protect Privacy Rights: Explicitly prohibit the use of generative AI for bulk domestic surveillance of U.S. citizens. -- Mandate Human Oversight: Support legislation requiring a meaningful "human-in-the-loop" for any AI-integrated lethal targeting system. -- Protect Appropriations: Ensure that defense funding is not used to retaliate against contractors for their commitment to AI safety and preventing illegal domestic surveillance. We must lead the world in AI by demonstrating that our technology is bound by law and ethics, not by removing the very guardrails that protect American citizens. National security is not a justification for abandoning the constitutional rights of Americans or the basic safety standards of our technology.