OpenAI

Anthropic Lets Claude Opus 4 Models End Abusive Chats in Rare Cases

The company is piloting the safeguard as part of its model welfare research to reduce exposure to extreme misuse.