OpenAI, Anthropic face ethical dilemmas in Pentagon AI talks
Business
OpenAI is negotiating with the Pentagon to bring its AI into classified systems—and is seeking to negotiate terms that would include technical safeguards and a "safety stack."
Meanwhile, rival Anthropic got a tough ultimatum: drop safeguards on its Claude model or risk losing a $200 million defense deal.
The ethical tightrope
Both companies are drawing clear lines—they don't want their AI used for mass surveillance or autonomous weapons.
OpenAI's Sam Altman even told staff the military won't force them to cross those lines.
It's a real test of how tech giants handle ethics when big government money is on the table.