Overview

• Anthropic revoked OpenAI engineers’ access to its Claude Code development tool for violating its terms of service prohibiting competing AI training.
• OpenAI expressed dismay over the move, saying its engineers were only evaluating and comparing models.
• ChatGPT now recognizes signs of mental or emotional distress and prompts users to take breaks during extended sessions.
• OpenAI collaborated with more than 90 global experts and industry groups to develop evidence-based guidelines for sensitive conversations.
• Apple has activated its “Answers, Knowledge and Information” division under Robby Walker to build an in-house chatbot, and OpenAI’s GPT-5 remains on track for a potential release by the end of August.