From Athlete to Avatar: How Creator Likeness Gets Licensed for AI Games and Fitness Apps

Quick answer: A growing category of licensing deals lets game studios, fitness apps, and interactive platforms build AI-generated avatars around a real creator or athlete's name, image, likeness, and captured movement — without that person needing to perform the actual content themselves. Done properly, it's a licensing deal like any other, built on consent and compensation. Done without carefully checking who actually holds the right to negotiate those rights, it's already produced real, high-profile legal disputes — including a lawsuit against a major video game publisher accused of negotiating directly with athletes in a way that bypassed an agency's existing group-licensing contracts.

The mechanics: what "licensing your likeness into a game" actually means

The basic pipeline looks like this: a company wants to build an interactive product — a video game character, a fitness app avatar, a virtual trainer — around a real, recognizable person, without needing that person to be physically present for every piece of content the product will ever generate. To do that, they license the person's name, image, and likeness rights, then typically use motion capture technology to record how that person actually moves, along with reference material for how they look and sound. AI systems then use that captured data to generate new content — new workouts, new game sequences, new interactions — that weren't individually performed or scripted by the person being represented, but are built from their real, licensed movement and appearance data.

This is a meaningfully different arrangement than a traditional celebrity endorsement, where a well-known person appears in content they actually performed. Here, the person licenses the raw material — their likeness and movement — and the company generates variable, ongoing content from it afterward, often without further direct involvement from the person themselves.

Where this is already happening

Fitness and gaming apps aimed at younger and general-consumer audiences have been an early, active category for this kind of licensing — approaching athletes, trainers, and content creators to license their name, image, likeness, and captured movement so an app can generate AI-driven workouts or interactive content featuring a recognizable, trusted figure at scale, across many more sessions and variations than that person could realistically perform live.

The appeal for the licensing company is straightforward: a recognizable, trusted persona drives engagement and credibility in a way a generic avatar doesn't. The appeal for the person licensing their likeness is a revenue stream that doesn't require ongoing production work — once the capture is done, the licensed content can be generated indefinitely without additional time from the person themselves, which is a fundamentally different economic model than trading hours for content.

When it goes wrong: the EA lawsuit

The category isn't without real, high-stakes disputes over exactly who controls the negotiation on a creator's or athlete's behalf. In 2023, The Brandr Group — a company that holds group name, image, and likeness licensing contracts with dozens of colleges — sued Electronic Arts, alleging the publisher unlawfully circumvented those existing group-licensing agreements by negotiating name, image, and likeness deals directly with individual college football players (reportedly around $500 per athlete) for an upcoming college football video game, rather than negotiating through The Brandr Group as required by its contracts with the schools (Law360). EA was not accused of using athletes' likenesses for free — the athletes were being paid directly. The dispute centers on a question this entire category runs into repeatedly: when an agency or licensing group holds a broader contract covering a group of athletes or creators, does a company have to negotiate through that group, or can it negotiate individually with the people covered by it?

The case illustrates a tension well beyond sports gaming specifically: as more creators and athletes route their name, image, and likeness rights through agencies or group-licensing arrangements, companies building AI-driven or motion-capture-based products have to be far more deliberate about whether an existing group agreement actually gives them the right to negotiate directly with individuals covered by it, or whether doing so exposes them to a claim from the agency holding those contracts.

What creators should ask before licensing likeness into an AI product

Is the capture reusable beyond this specific product? Motion capture and likeness data, once collected, can technically be reused across multiple products unless the license specifically limits that. Scope matters as much here as in any other licensing deal — confirm whether the license is limited to one app or game, or whether it extends to future, unspecified products built by the same company.

Who controls how the AI-generated output is used? Because the licensing company is generating new content after the fact — content the licensor never individually reviews or approves in real time — it's worth negotiating some form of approval right or content guideline upfront, rather than after something objectionable has already been generated and published.

What happens to the captured data if the deal ends or the company changes hands? Motion capture and likeness data are valuable, durable assets. A contract should specify what happens to that data — deletion, continued licensed use, or transfer restrictions — if the relationship ends or the company is acquired.

Does compensation scale with usage? Because AI-generated content from a single capture session can be reproduced far more times than a traditional appearance or endorsement, a flat one-time fee may significantly undervalue the actual scale of use. A revenue share or usage-based royalty structure is often a more appropriate match for how this kind of licensing actually gets deployed.

Key takeaways

Licensing likeness into AI-driven games or fitness apps typically combines a name/image/likeness license with motion-capture data, letting a company generate ongoing content without the person's continued direct involvement.
The EA/Brandr Group lawsuit illustrates a recurring dispute in this category: whether a company can negotiate name, image, and likeness deals directly with individual athletes or creators who are already covered by an agency's group-licensing agreement, or whether that requires going through the agency holding those contracts.
Creators evaluating these deals should clarify reusability limits, approval rights over AI-generated output, data handling after the deal ends, and whether compensation scales with actual usage.
This is a fast-growing licensing category precisely because it decouples a creator's income from their ongoing time and production work.

FAQ

Can a video game legally use an athlete's likeness without their direct permission?

Generally, no — using a real, identifiable person's name, image, or likeness commercially requires a license or a legitimate claim to one. The Brandr Group's lawsuit against Electronic Arts illustrates a related wrinkle: even when a company is paying and clearing rights directly with individual athletes, doing so can still trigger a legal dispute if an agency holding a broader group-licensing agreement argues that its contract should have controlled the negotiation instead.

How does motion capture licensing work for AI-generated avatars?

A company records how a real person actually moves using motion capture technology, combined with reference material for their appearance and voice. That captured data becomes the basis for an AI system to generate new content — new workouts, game sequences, or interactions — built from the person's real, licensed movement without requiring them to perform each new piece of content individually.

Should compensation for likeness licensing be a flat fee or based on usage?

Because AI-generated content from a single licensed capture can be reproduced far more extensively than a traditional appearance, a usage-based or revenue-share structure often better reflects the actual scale of use than a one-time flat fee.