OpenAI CEO Sam Altman met with U.S. lawmakers to address scrutiny over the company's new Department of Defense contract, focusing on AI use in warfare and surveillance.
OpenAI's Defense Contract and Rival Blacklist
- OpenAI secured a deal with the DOD late last month, permitting AI use for "all lawful purposes."
- This followed rival Anthropic's blacklisting by Defense Secretary Pete Hegseth over national security supply-chain risks.
- Anthropic's DOD negotiations stalled due to disputes: the military sought unrestricted model access, while Anthropic demanded limits on autonomous weapons and domestic surveillance.
Lawmakers' Ethical Concerns and Discussions
- Altman met with lawmakers in Washington D.C., including Senator Mark Kelly (D-Ariz.), who raised "serious questions" about OpenAI's warfare approach and the DOD contract.
- Kelly said discussions detailed surveillance and AI integration into a "kill chain," calling it a "good discussion."
- Kelly is collaborating with senators to draft legislation establishing guardrails for DOD-AI contracts, emphasizing congressional oversight.
OpenAI's Safety Principles and Contract Terms
- Altman highlighted OpenAI's core safety principles: prohibitions on fully autonomous weapons and domestic mass surveillance.
- The DOD contract reportedly embeds these principles, with OpenAI confident that existing laws and internal safeguards prevent misuse.
- Altman expressed support for U.S. government partnerships but disagreed with the DOD's Anthropic blacklist decision.
Push for Legislative Guardrails
- Senator Kelly is advancing legislation to set boundaries for defense AI contracts, noting technology's rapid pace outstrips legislative speed.
- The effort aims to ensure ethical AI deployment in national security, amid growing industry-government tensions.