Anthropic turned down a reported $500 million US government contract for human targeting and surveillance. OpenAI took it. Mo argues this is the clearest ethical signal in the industry — and that every time you choose which AI to use, you're casting a vote.