“Your face will not ever make it to production,” xAI told staff — but concerns remain.
Elon Musk’s AI firm, xAI, is facing internal unrest after over 200 employees took part in “Skippy,” a training project aimed at teaching its chatbot Grok how to interpret human emotions using facial videos. Internal documents seen by Business Insider show that staff were asked to record themselves making expressions and engaging in mock conversations, sparking concerns about data use and privacy.
“Your face will not ever make it to production,” a lead engineer assured during the project’s kickoff, adding the goal was to “teach Grok what a face is.”
Despite promises of internal-only use, consent forms granted xAI perpetual rights to workers’ likenesses, including for commercial purposes.
Tension grew after xAI released lifelike avatars, Ani and Rudi, which users quickly manipulated into sexually explicit or violent behavior.
One employee voiced concerns: “Could my face be used to say something I never said?”
xAI declined comment. Meanwhile, the company continues avatar development — including a new Grok companion inspired by Twilight’s Edward Cullen.