“Both OpenAI and Anthropic have published reports claiming to have detected Chinese state-affiliated hackers using their models for malicious activities.”