How EA used AI to scan ~11K players' photos to create 3D avatars for its college football game; the NCAA first let players to sell their likeness rights in 2021
Electronic Arts used new tech to scan photos for its videogame after securing players' likeness rights for the first time X: @saraheneedleman X: Sarah E. Needleman / @saraheneedleman : To release the college-football game this month, EA relied on artificial-intelligence technology it began developing about four years ago, before the NCAA set a new policy allowing players to sell their likeness rights in 2021 https://www.wsj.com/...
Context & Ripple Effects
EA’s college-football return combines two inputs that have historically been separate: scalable synthetic-image tooling and permission to use athlete identities. Earlier coverage of AI firms selling realistic generated faces illustrates the underlying capability now being applied to a licensed entertainment product: realistic AI-generated faces.
For EA, rights management is already a strategic issue in sports games: its football franchise became FC 24 after EA and FIFA ended their licensing relationship. The NCAA’s policy shift gave EA a route to assemble athlete permissions for this title rather than relying on an unlicensed approximation.
First-order effects
- EA can populate its college-football game with 3D representations of roughly 11,000 authorized players, using photo-scanning AI instead of a wholly manual avatar-production process.
- College athletes whose likeness rights EA secured become represented in the game under the post-2021 NCAA policy framework.
Second-order effects
- Sports-game publishers seeking comparable authenticity face a two-part requirement: obtain usable likeness rights and build a production pipeline that can translate supplied images into game-ready assets at scale.
- The value of licensing operations rises alongside AI asset generation: better generation tools do not remove the need to clear the underlying identities.
Third-order effects
- If this model is repeated, sports-game competition will increasingly turn on control of rights networks and consent workflows as much as on graphics and gameplay technology.
- The case reinforces a durable boundary for commercial generative AI: scaling depiction of real people depends on governance and authorization, not merely model capability.
The trend: Licensed AI-assisted asset creation is turning likeness governance into a core production advantage for digital entertainment.