Overview

• Project Skippy enlisted over 200 AI tutors to record 15- to 30-minute videos of their faces and interactions to teach Grok to interpret human emotions.
• Employees were assured that their faces would never appear in production but consent forms granted xAI perpetual rights to use their likeness in training and commercial offerings.
• Dozens of tutors flagged unclear usage boundaries and potential misuse of biometric data, and several formally opted out of the initiative.
• The facial-data effort follows controversies over Grok’s antisemitic outbursts and the release of provocative avatars such as Ani and Rudi.
• Despite detailed documentation of employee objections, xAI has remained silent in response to requests for comment on Project Skippy.