Elon Musk said Tuesday his social media platform X is working to fix its AI chatbot, Grok, after it sided with rival tech chief and OpenAI CEO Sam Altman in an ongoing dispute with Musk, who threatened to sue Apple on Monday night over allegations it favors OpenAI in its app store.

More than 370,000 Grok chats have been indexed by search engines, exposing sensitive queries that include intimate medical and psychological questions, business details, and at least one password.

A chatbot run by Elon Musk’s X told users how to assassinate its billionaire chief executive. Grok, which can be accessed on Mr Musk’s social network, has instructed users how to make bombs and kill themselves, according to hundreds of thousands of leaked chats were exposed on Google.