Protecting Athletes’ Image Rights in the Age of AI Deepfakes

Bruno Fernandes has recently become the latest target of an AI-generated deepfake advertising campaign. The Vietnamese-registered sportsbook and casino QH88 built an entire website around a fabricated partnership with the Portuguese international, devoting considerable resources to producing footage showing the player apparently signing an "ambassadorial contract" with QH88 representatives at Old Trafford.

A Growing Pattern

This is far from an isolated incident. It is part of a growing pattern of AI being used to fabricate endorsements of the world’s biggest sporting stars. It is reported that at least 10,000 deepfake advertisements impersonating football players have been posted online in the last 12 months. The 2026 World Cup only accelerated this activity, with stars such as Cristiano Ronaldo understood to have been targeted by an average of more than eight fake AI-generated adverts a day during the tournament.

The Commercial Stakes 

Beyond the immediate reputational damage, this scale of exploitation poses a direct commercial threat to players. Unauthorised deepfakes risk undermining the exclusivity that underpins the image rights structures that athletes rely on to monetise their identity, while jeopardising both existing sponsorship arrangements and their ability to secure future endorsement deals. 

The Current Legal Landscape 

Under UK law, the use of deepfakes to facilitate an already criminal act is illegal, such as deceiving facial recognition software to fraudulently access online banking. However, when deepfakes are used for purposes that are not independently criminal, such as fabricating a commercial endorsement, the legal position is less clear. The starting point under UK law is that an individual does not have automatic rights over their own personality or image rights, meaning protection for athletes must be carefully pieced together using a variety of legal doctrines. 

Using Trade Marks to Protect Athletes’ Personal Brands

The rise of deepfakes has pushed athletes to think more deliberately about which parts of their identity are worth protecting, with trade mark law proving to be a central part of that strategy. For example, Cole Palmer has secured protection for his signature goal celebration as a registered motion mark following a 2018 amendment removing the Trade Marks Act's "graphical representation" requirement. 

Luke Littler's recent application follows in that spirit, seeking to register his own likeness on top of existing protection for his name and nickname. Taken together, the two filings suggest a wider shift where athletes and their advisers are no longer treating a name as sufficient protection on its own, instead mapping out every distinctive element of a personal brand as something worth registering individually. 

Trade marks won’t prevent a deepfake from being made in the first place, but they strengthen the position of athletes when AI-generated content crosses into commercial territory or creates the false impression of a genuine endorsement. Infringement requires unauthorised use of a mark "in the course of trade," so a deepfake that reproduces a registered gesture or likeness element to promote a product or service gives rise to a comparatively direct infringement claim. 

However, this route does not come without limitation. The UK Supreme Court decision of SkyKick UK Ltd v Sky Ltd [2024] UKSC 36 confirmed that a trade mark registration can be partially invalidated for bad faith where an applicant had no genuine intention to use the mark across the full breadth of the specification claimed. Following that decision, the UK Intellectual Property Office issued Practice Amendment Notice 1/25, under which examiners may now raise bad faith objections during the examination of applications with specifications that are “manifestly and self-evidently broad.” Athletes seeking to protect themselves against deepfakes using trade marks must therefore balance the breadth of their protection against the very real risk of that protection being challenged, whether by the IPO at examination stage or by third parties in opposition or invalidation proceedings. 

Where Trade Marks Fall Short – Alternative Causes of Action

Registering an athlete’s likeness remains a new and developing concept, so there are alternative causes of action for athletes targeted by false advertisements who are not under the protection of a mark. Although it did not deal with the issue of a deepfake, the case of Irvine v Talksport Ltd [2003] EWCA Civ 423 gives reassurance that English courts are likely to consider such content as passing off. That decision established the principle that using a person’s image without consent to create a false impression of endorsement can amount to a misrepresentation actionable under the law of passing off. 

To make a successful passing off claim, an athlete will need to prove three elements: goodwill, misrepresentation and damage. Goodwill requires the athlete to show that their name or image carries sufficient commercial value to influence a customer’s purchasing decisions. Once established, misrepresentation requires the deepfake to be sufficiently convincing that consumers believe the endorsement is genuine, a threshold which is becoming easier to meet as AI technology advances. Finally, the athlete must prove that this misrepresentation caused them damage. This can create difficulty, as athletes ideally need to be able to point to clear detriment such as loss of revenue from image rights agreements or sponsorship deals. 

Data Protection and Defamation 

Athletes could alternatively turn to data protection legislation to combat false advertisements. Creating a deepfake involves processing biometric data, arguably a “special category” of data under UK GDPR, needing consent or manifest public availability as a lawful basis. However, no English case has yet tested this route against a creator of a deepfake. 

In addition, a defamation claim may be available for the athlete where the deepfake causes serious reputational harm. This finds precedent in Tolley v Fry, where an amateur golfer’s unauthorised use in a chocolate advertisement was held capable of being defamatory, as it falsely implied he had commercially exploited his reputation for payment, risking his amateur status. This reasoning applies to deepfake advertising, with the reputational harm lying in the implication that the athlete chose to endorse the brand in question. This can be particularly damaging where that brand carries obvious reputational risk. However, there is currently no settled English case law on the application of defamation law to deepfakes involving athletes.

The Jurisdictional Challenge

Although athletes may have recourse through the causes of action discussed above, protecting against deepfakes used in advertisements is far less certain where no registered trade mark exists. This uncertainty is compounded by two separate obstacles. The company responsible for creating or commissioning the deepfake may be based anywhere in the world, making it difficult to identify or pursue directly. The platforms hosting and disseminating the content are themselves decentralised, and are often slow to act on takedown requests. Together, these challenges make the enforcement of traditional remedies, such as injunctions, considerably more difficult.

The EU AI Act offers a partial response to these enforcement challenges. Under Article 50, which took effect on 2 August 2026, providers of AI systems that generate synthetic content must mark their outputs in a machine-readable format so that they are detectable as artificially generated. Deployers, in turn, are required to disclose to users where content constitutes a deepfake. In practice, however, these obligations are unlikely to be observed by those knowingly creating deceptive content in the first place, limiting their effectiveness in this context.

A Commercial Opportunity? 

For all the risk deepfakes pose, the same underlying technology also opens up a genuine commercial opportunity and some individuals are already putting it to their own advantage. For example, Neymar Jr has licensed his likeness to FlareFlow, an international micro drama platform owned by digital publisher COL Group. As part of this deal, Neymar’s AI-generated likeness will feature in a 16-title franchise of AI-assisted vertical series (i.e. in portrait orientation), led by the title “The Way Back to Glory.” Similarly, NBC used AI to recreate sportscaster Al Michaels' voice for personalised highlight packages during its coverage of the 2024 Paris Olympics on Peacock, with Michaels compensated for, and having approved, the use of his likeness.

Both arrangements illustrate the growing scope for athletes to draft carefully structured endorsement agreements that licence their likeness for passive, ongoing revenue. This enables them to turn the same underlying technology that threatens to undermine an athlete's commercial identity into a controlled and monetisable asset instead.

Practical Steps for Athletes and Their Advisers 

1. Review existing image rights agreements 

Many current agreements were drafted before generative AI made convincing content easy to produce and may not adequately address AI-generated use at all. 

2. Introduce express “synthetic rights” clauses 

Agreements should explicitly address the use of biometric data and set clear approval rights before any AI-generated content bearing an athlete’s likeness is published. 

3. Consider strategic trade mark registration 

Registering narrowly defined elements of an athlete’s brand such as a name, gesture or likeness can offer a direct route to enforcement, provided filings are structured carefully. 

4. Build in takedown protocols

Due to the speed at which deepfake content spreads, agreements between athletes, agents and clubs should formalise responsibility for monitoring platforms and responding quickly to unauthorised content.

5. Explore licensing opportunities 

Rather than treating AI purely as a threat, athletes can decide to licence their likeness on their own terms, turning the same technology into a controlled commercial asset provided the agreement is drafted with real precision over what exactly is being licensed and to whom. 

In an environment where a single deepfake can spread rapidly across the globe, protecting an athlete's commercial identity now requires a considered, proactive legal strategy built before the damage is done. With experience spanning trade mark strategy, image rights protection and commercial licensing, organisations such as Level are positioned to help athletes and their advisers navigate a legal landscape that is evolving as quickly as the technology itself

As a leading sports and media firm with a specialist IP team, including trade mark expert Nick White, Level is on hand to assist with any aspect of IP protection or strategy advice discussed in this article.