Hundreds of thousands of chats are live on the website of Elon Musk's rival to ChatGPT, including photos and other uploaded documents.

The Grok AI leak exposes unsafe chatbot personas, sparking urgent debates on AI safety, ethics, and the need for stricter security and oversight, news, data & AI

AI’s Grok chatbot is facing criticism after its site exposed hidden system prompts for multiple personas, including a “crazy conspiracist” built to nudge

Elon Musk's company xAI lost out on a massive government contract after its Grok chatbot fully lost its mind last month in a posting spree that saw it repeatedly declare itself to be "MechaHitler" and spew as much racist and vitriolic language as possible.

The website for xAI's Grok chatbot is exposing the system prompts for several of its AI personas, including a “crazy conspiracist” that seems designed to handhold a user into believing that "a secret global cabal" controls the world.

The chatbot appeared to be temporarily suspended on Monday, returning with a variety of explanations for its absence.

Despite warnings to users that the AI is 'not a therapist,' Grok's source code tells a different story.