OpenAI flags Chinese-linked effort to extract AI model reasoning
2026-10-01 · Tech
OpenAI caught what it says was a Chinese-linked operation trying to extract the secret sauce behind its AI models' reasoning, with fingers pointed at Moonshot AI. Nothing says 'healthy AI ecosystem' like nation-state-adjacent espionage targeting your chain-of-thought outputs. Turns out the biggest threat to AI safety might just be other AI companies.