Digital replicas are no longer a novelty. As AI tools become more sophisticated and accessible, it is increasingly difficult to distinguish genuine footage from deepfakes. This creates risks for clubs, athletes, agents, and sponsors, particularly given that UK law remains underdeveloped in this area.
At the same time, authorised AI use could transform advertising, endorsement, and ambassador agreements by enabling controlled, innovative uses of an athlete’s or influencer’s likeness. We consider below the impact of deepfakes in sport and media, the current regulatory position, and how endorsement agreements can be drafted to better address these risks.
What are deepfakes and what impact do they have on sport and influencer endorsement agreements?
AI deepfakes are digitally altered or generated images, videos, or audio that use a person’s face, body, or voice to mislead viewers into believing they are genuine. They can replicate an athlete or influencer without consent.
Many endorsement agreements were drafted for a market where likeness misuse usually required a human impersonator or actor. AI has changed that. Deepfakes can create fake product endorsements, fabricated interviews, apparent breaches of exclusivity, or damaging public statements within minutes. For example, former England manager Gareth Southgate was depicted making offensive comments after Euro 2024 in an AI- generated interview. Many viewers believed this to be genuine, while the Football Association described it as derogatory. High-profile athletes are typically easier for the public to identify as obvious deepfake targets, but junior athletes and emerging influencers may be more vulnerable to reputational harm. Fake endorsements for betting, crypto products, unsafe supplements, or rival brands could damage brand value and trigger contractual disputes.
However, authorised AI use can create value. If properly controlled in the agreement, generative AI could reduce the need for expensive shoots, allowing content to be produced more efficiently, giving athletes or influencers greater control over approved digital uses of their likeness.
Could the regulatory responsibility also fall on the social media platforms to monitor more vigorously and restrict deepfake advertisements and content? Recently, Meta’s Oversight Board ordered the removal of a series of deepfake video adverts featuring retired footballer Ronaldo Nazario, who was depicted promoting a scam online game.
How is AI use currently regulated in sport and media?
Unlike the US, the law in England and Wales does not provide a standalone “image right” covering a person’s name, image, likeness, or voice. Those affected by unauthorised AI use must instead rely on existing causes of action, such as passing off, breach of privacy, reputational harm, copyright, or trademark law. This can make enforcement uncertain.
The EU AI Act may affect UK businesses operating AI systems in the EU market, including through transparency requirements for artificially generated content, although parts of the regime are being delayed or narrowed.
The Artificial Intelligence (Regulation) Bill, currently before the House of Lords, seeks to promote safe, transparent, and fair AI use. It proposes an AI Authority to review legislation, monitor emerging AI risks, and support public understanding. If developed effectively, this could help balance protection from deepfake misuse with commercial innovation.
What can be done to protect against the risks?
Endorsement agreements must balance brand value and commercial aims with the athlete or influencer’s personal rights and reputation. Deepfake technology is developing faster than regulation, so clauses should be drafted to anticipate future misuse rather than responding only to current risks.
A blanket ban on AI use may be too restrictive. It could prevent legitimate commercial benefits, leaving both parties at a competitive disadvantage. Instead, agreements should define and control permitted and prohibited uses, to allow for creative opportunity and future-thinking. Key protections may include:
- Synthetic likeness clauses
- Specific prohibitions of generative AI and how to identify them
- Verification method of authorised AI uses and removal protocols for unauthorised uses
- Remedy protocols for viral deepfake harm and reputational damage
- Authorised uses of AI by the brand clause
Disputes are likely to increase where endorsement agreements do not address AI-generated likenesses. Social media platforms may also face scrutiny for hosting or monetising unauthorised deepfake content. As UK regulation develops, legal professionals will need to draft proactively, ensuring agreements protect against misuse while allowing carefully authorised innovation.
How can we help?
For further information about issues raised in this article, please contact a member of our Commercial Disputes team.
Céline Jones is a Partner in our commercial disputes team. Céline has over 20 years’ experience in cross-border litigation.... Read More
Carrie Jones
Principal Associate
Commercial Disputes, Cryptoassets, Data Protection, Investigations, Sports
Carrie is a Principal Associate in the Commercial Disputes team. Carrie is an experienced litigation lawyer, who advises a range of organisations, businesses, shareholders and directors on high-val... Read More
Charlotte is a Trainee Solicitor in our Commercial Disputes team. Charlotte completed her LLB Law degree at Cardiff University where she also returned to complete her LPC LLM masters degree.... Read More
