Anthropic is revolutionary a concept known as “constitutional AI” – that's correctly a way to make an LLM safer by making sure it “gains” values via a constitution presented. None especially targeting ChatGPT. But OpenAI is involved in at the least a person lawsuit which includes implications for AI methods https://ai-bot29723.blogminds.com/the-smart-trick-of-ai-bot-that-nobody-is-discussing-28863122