xAI allegedly required employees to submit biometric data like facial scans and voice recordings to help train its anime-style chatbot, Ani.
According to The Wall Street Journal, the data collection was part of a confidential internal program known as Project Skippy.
Ani was introduced back in July as part of the $30-a-month SuperGrok subscriber tier on X. The female avatar was promoted as a more human-like companion powered by Grok’s AI.
According to the Journal, xAI asked designated “AI tutors” to sign release forms granting the company broad, royalty-free rights to their likeness and voice. Some staff reportedly pushed back, citing concerns over potential deepfake misuse and the chatbot’s sexual tone.
Project Skippy and employee concerns
The biometric data release forms granted xAI perpetual, worldwide rights to use, reproduce, and distribute employee faces and voices. Staff were told the data would improve Ani’s realism and help train future Grok-linked companions.
Employees who objected to the request were told it was a job requirement tied to xAI’s mission. Some were reportedly uncomfortable with Ani’s design, which some staff viewed as resembling a waifu character.
In fact, right when she launched, fans immediately drew comparisons to other women, both real and fictional, believing she was designed with Grimes or Death Note’s Misa Amane in mind.
Grimes, who was in a relationship with Musk for a few years, hinted that Ani was modeled off her with the release of the music video for her song ‘Artificial Angels.’ The video featured Ani throughout and included scenes with the singer dancing with the AI chatbot.
