US and China to Discuss AI Hotline, OpenAI Employee Account Hacked

Beijing and Washington are set to discuss an AI hotline, focusing on governance amidst rapid AI advancements. Separately, a small team of hackers reportedly used Anthropic's Claude Opus 5 to compromise an OpenAI employee's ChatGPT account and access sensitive code. Researchers at Google Cloud and the University of Waterloo also developed 'ScientistTwo,' a multi-agent AI system capable of conducting parts of the scientific method autonomously.