Deepfakes, Satire and Copyright. Where Does UK Law Draw the Line?

Artificially manipulated media have outpaced the general public’s understanding and, at times, the law itself. In the UK, deepfakes and other synthetically generated content exist at the complex intersection of copyright, privacy, data protection, defamation, and criminal law. While satire and parody are protected forms of expression, the notion of ‘it’s just a joke’ does not serve as a legal shield from infringement. This underscores the necessity for a comprehensive understanding of the intricate and occasionally perplexing landscape of the law in this area.

The “fair dealing” exception and its limitations

UK copyright law contains a narrow safety valve for humour and comment. Since 2014, the Copyright, Designs and Patents Act 1988 has included an exception allowing fair dealings for parody or caricature. In simple terms, this means you can use a limited amount of someone else’s work without permission, provided it is genuinely used for parody and is fair. The UK government’s guidance underscores that ‘only to the extent that is fair dealing is permitted’, meaning there is no blanket wholesale. The concept of fairness, however, is contextually dependent, and factors that determine whether something is fair or unfair are not always objective.

Typically, it is considered whether you have used more of the original than necessary to transform or, in other words, conjure up your new work and whether the new work competes with the original work or acts as a substitute for it. It is also considered whether the new work is transformative, adding new expressions or meanings that the original cannot portray in the same way, and whether it harms the market of the original work. Despite seeming relatively easy to follow, there is never an objectively right or wrong decision, as people have different standards as to what is transformative.

Moral Rights and Attribution

It is essential to note that even if your use is considered ‘fair’ and ‘transformative’, authors will always retain the moral rights to their work. This means that they have the right not to have their work subjected to derogatory treatment. Derogatory treatment can include altering the work in a way that is damaging to the author’s reputation or presenting the work in a context that is offensive to the author’s beliefs. What this means is that parody versions do not always overpower those claims and are not a blanket excuse for unfair or non-transformative content.

When it comes to deepfakes, the potential for legal breaches is significant. A single deepfake video could infringe upon multiple rights, including musical, audiovisual, photographic copyrights, and notably, performance rights. Even if an individual is cleared of infringement on one of these potential breaches, it does not absolve them of all. This underscores the urgency of compliance and the need for a thorough understanding of the legal landscape.

The UK law allows an element of satire, but not complete copying where the “parody” republishes substantial parts or chunks of the original work or acts as a full substitute; the “parody” falls outside the exception by law.

Image rights

Unlike some jurisdictions, the UK does not recognise a freestanding ‘image right’ which was established in Rihanna v Topshop 2015, where the London High Court of Justice held that Rihanna could not claim a monopoly over her likeness. Nevertheless, Rihanna did succeed, not by claiming ‘image rights’ but by passing off, a legal concept that refers to the misrepresentation of a product or service as being endorsed by a celebrity or other well-known figure. This judgement is often cited as one of the exemplar cases which established that there are no general image rights, but passing off may be sufficient as a method of claim when there is a false suggestion of an endorsement.

When it comes to deepfakes, the same logic can be applied. If a generated, fake video or clip implies a celebrity or even a brand has endorsed a product or message when they have not, the creator of the video risks a passing off claim. Even when the creator is not selling anything, brand style or brand misrepresentation in promotional content can also be actionable if the goodwill has been damaged and shown.

Separate from deepfakes, it is also important to mention that individuals themselves can rely on the argument of misuse of private information, a privacy tort established through the Human Rights Act, balancing Article 8 and 10. Similarly, an individual can also claim using data protection law, which I explore further below, to protect themselves when personal data has been processed and used.  

Data protection and biometrics

Suppose a deepfake is created and uses or derives from a real person’s image, voice or other identifiable factors. In that case, the creator is likely acting in a manner classed as the processing of personal data under the UK GDPR and the Data Protection Act 2018. The Information Commissioner’s Office, otherwise known as the ICO, has specific guidance for artificial intelligence, emphasising the requirement of lawful basis, transparency, data minimisation and overall fairness. If a deepfake includes an individual, attributes, or creates a biometric identity/template. In that case, it is considered high risk and will often require a robust assessment to align with and follow the outlines stated through a Data Protection Impact Assessment.

Special Biometric category data

For creators, the use of biometric data, which uniquely identifies a person, is a matter of special consideration. It falls under a special biometric category that requires additional conditions for processing. This underscores the importance of obtaining explicit consent and maintaining strong governance when using real people’s data in ways they might not expect.

Defamation and malicious falsehoods

A deepfake that causes serious harm to an individual or a company’s reputation can be considered defamatory. The Defamation Act 2013 states that “a statement is not defamatory unless its publication has caused or is likely to cause serious harm.” This threshold filters out trivial and obvious jokes, but not realistic and potentially malicious manipulations of a system that reasonable viewers may struggle to decipher.

There are some defences available for a defamation claim, those being truth, honest opinion and a publication on a matter of public interest. However, these are not reliable. If the content is false but not necessarily reputation-damaging, and still causes measurable economic loss, a claimant might consider malicious falsehood. To protect yourself if you are engaging in this activity, avoid ambiguity, as highly convincing deepfakes lack clear satire signals where the viewers could be misled.

Harassment and misuse of private information

Deepfakes that target a specific person, such as leads or individuals in campaigns, can trigger the Protection from Harassment Act 1997, which provides the individual with both criminal offences and a civil course of action via injunctions and damages. The term “Harassment” is defined as a course of conduct causing alarm or distress, and online conduct and publishing of harassing material purporting to originate from a person can contribute to liability. When a deepfake intrudes into an individual’s private life, fabricating intimate scenarios or faking financial reports, it is considered a misuse of private information. Privacy claims are often far more direct than copyright claims.

Online Safety Act and explicit deepfakes

The Online Safety Act 2023, otherwise known as OSA, introduced new communications offences that came into effect on 31 January 2024. This included offences around false communications intended to cause harm and intimate image abuse. The OSA amended the Sexual Offences Act 2003 so that sharing or threatening to share intimate images without consent, including deepfake pornography, is considered a serious criminal offence. In April 2024, the government announced its plans to go further by criminalising the creation of sexually explicit deepfakes of adults, which planned to recognise the gaps where creation itself causes the harm even without distributing the deepfakes.

Satire or Deception?

One of the most essential elements to consider when reviewing whether something is satire or deceptive is the presentation. Signposting that your work is a parody with framing devices such as overt captions or explicit disclaimers that an average viewer would recognise as satire. It is also essential to use the minimum necessary and use just enough of the original work to achieve the satirical point, which is critical, as overusing the original work undermines the fairness principle. In addition, it is crucial to avoid market substitution.

What this means is that if your work competes with the original, for example, uploading or re-uploading a whole video with minor gaps, will cause problems and will infringe upon the creators’ rights. Another essential element to avoid is implied endorsements, meaning do not use brand logos or taglines. It is crucial to gain explicit consent for satire work that involves intimate, biometric or reputationally risky scenarios, as ethically the work may be brought into question as to whether it breaches any data protection or privacy laws.

Checklist for creators and publishers

1. Purpose: Check your work. Does the use genuinely parody and satire, or is it used as a vehicle for the distribution of the original work without their permission?
2. Quality Check: Have you incorporated the minimum amount of the original work into your new job, ensuring it is not redistributed with little or no changes?
3. Presentation Check: Could a reasonable viewer decide about the reality or believe that the work has been endorsed by the individual in the new job? If it is unclear, it is important to add clear signals or re-edit the work to avoid a defamation claim or passing off.
4. Privacy and Data Check: Is there any chance that the work could be perceived to be processing personal or biometric data? It is essential to identify the law and consider obtaining explicit consent whilst being fully transparent.
5. Check the type of content: If the work involves anything intimate or sexual, it runs the risk of being a criminal offence, so it is best to stay clear of anything about this.
6. Could it be considered as harassment? Is there a pattern to your work or repeated targeting or humiliation of an individual, which could lead to a criminal or civil issue, even if each post is small?
7. Have you kept records? It is essential and may prove very useful later if any legal issues arise. Keep note of intent, sources and any edits you make.

Conclusion

In conclusion, the UK law protects parody and satire, but only when they fall within the fair dealing limits and are shown without deception. If a deepfake substitutes for the original work, thereby implying endorsement without the endorsee’s consent, or if it intrudes on privacy, causing serious reputational harm, harassment, or the dissemination of intimate imagery, a civil and criminal issue will likely arise.