Overview

• xAI launched the Skippy project in April, asking over 200 AI tutors to record 15- to 30-minute videos of themselves to teach Grok how to interpret human emotions.
• Participants were required to sign consent forms granting xAI perpetual rights to their facial data and likeness for training and commercial products, a provision that alarmed many.
• In internal Slack messages and meetings, tutors flagged privacy and ethical concerns and several chose to opt out of the initiative.
• Engineers assured participants that their videos would be used solely for training and would never appear in production, but employees remain skeptical.
• xAI has not commented on whether the Skippy footage contributed to its recently released Ani and Rudi avatars or on planned safeguards for the collected biometric information.