AI Training on Music

Michael Bywell

michael@newmedia.law

michael@bywellconsulting.com

In this article Michael Bywell, independent consultant and Partner at New Media LLP summarises the direction of travel on AI and music copyright issues.

Introduction

The questions here concern whether recordings, lyrics and compositions can be copied and processed to train generative AI models and then exploited without the permission of or payments to rightsholders. These issues are being tested in cases against music AI platforms in several jurisdictions including the US, Germany, Canada and Denmark. And in the UK, the government concluded a consultation earlier this year on how UK copyright law should respond to generative AI which is discussed below. The Getty case has also made waves and has relevance even though it is not a music industry case.

The scale of the issue

The actual scale is uncertain mainly because the training statistics are not publicly available nor readily disclosed. However, to give some idea, Suno admits to having trained its model on ‘tens of millions’ of audio files and, on usage, claims that more than 100m people have made music using Suno since it started up in 2022.

US

In the US, Universal, Warner and Sony (the ‘Big 3’ music companies) have all sued Suno and Udio. Suno and Udio are GenAI platforms that allow anyone to create full length music tracks from simple prompts. Some cases have settled and new licensing arrangements have started to appear. Universal and Warner (but not Sony) have settled with Udio; and Warner (but not Universal or Sony) has settled with Suno. We are also seeing artists bringing claims asking for more money of the back of these cases and deals.

Germany

In July 2026 a German court ruled against Suno in relation to the unauthorised use of six well known songs including two Boney M tracks and the song ‘Mambo No. 5.’ The case was brought by GEMA, the German collecting society on behalf of composers, lyricists, and producers. The Munich Regional Court found evidence of copying without permission at the training stage; memorisation inside the model meaning reproductions were stored on the model; and evidence of further infringements at the output stage. In relation to the latter, these infringements were supported by tests that GEMA had undertaken during which they entered open-ended prompts with song title, lyrics, and description of style, and received back recognisable musical elements of original works. Suno argued, unsuccessfully, that the model was learning mathematical patterns and relationships only but the court said that a recognisable form of a work could be generated so there was evidence of reproduction for copyright purposes. This was a robust and bold decision by the Munich court not least because it took it upon itself to consider both German and US copyright law. The US element was relevant to training that took place on US soil. There are reports that Suno is considering an appeal from that decision.

UK

The Getty Images v Stability AI case [2025] EWHC 2863 (Ch) is worth noting even though it is not a music case. It concerns alleged infringement by Stability AI of Getty’s rights over photos, videos, illustrations and digital artwork. The primary infringement claim (data scraping and copying to train the model) was dropped because there was no evidence that training had taken place in the UK. The secondary infringement claim relating to the output or commercialisation of scraped and copied data failed because there was no evidence that images were stored on the model. Technically speaking, the court held that the model’s numerical weights (numerical values representing patterns learned during training) were not themselves copies of the Getty images. So there was no infringement. There is an appeal which is due to be heard next month.

UK government consultation

On 18 March 2026 the UK government published its “Report on Copyright and Artificial Intelligence”. In short, a stocktake of how UK copyright law should respond to generative AI:

Copyright exception?

The report says there will be no copyright exception to allow training on models without the permission of rights holders. Originally the UK government had said there would be an exception coupled with an opt-out right for rights holders. However there was considerable push back with the creative community saying that opt out would not work and in truth meant ‘permission by default.’ It would be too hard for copyright owners (e.g., illustrators, photographers, musicians or writers) to monitor potential infringements and identify (i) that their work had been scraped, (ii) every GenAI developer using it, and (iii) register objections. Furthermore, that GenAI companies should be required to seek permission in the first instance and it was self-evidently wrong to allow those companies to utilise the content of others without payment and then use it to compete against those same rights holders. The government listened to those concerns and changed its mind.

Transparency on what is being used to train models

Reflecting the objections already highlighted above, the report concludes that there is a need for more transparency on what material is being used to train systems. Artists and rights holders support this because it makes it easier for them to discover whether their work has been used and enables them to object to unauthorised use, seek payment and negotiate licences where possible. The government is therefore looking at best practice for input transparency. Legislation has not been ruled out.

Licensing

The report recognises licensing as a potential mechanism for ensuring rights holders are paid when their work is used in GenAI models. The government does not propose intervening and instead will see how the market develops.

Digital replicas

This concerns AI’s ability to replicate someone's voice, face, likeness and identity as a performer - for example, where GenAI generates a completely new song using a synthetic voice that is recognisably that of a human artist. There may be no copying of an artist’s actual recordings or performances but one or more of the artist’s characteristics may have been used. For a singer, this could be very damaging because it is unauthorised use of their identity and could give rise to reputational and financial damage. It might be a terrible or politically motivated song that the original artist does not want to be associated with. The rogue track might also compete with the true artist and potentially deprive them of income.

The report recognises that UK law does not currently cover every situation where someone's voice or likeness is replicated without consent and because of this the government proposes exploring whether the UK should introduce a new “digital replica” or “personality” right.

Why are we not seeing a tidal wave of cases?

Given the amount of music said to have been scraped in recent years and the potentially adverse consequences for artists why is it that we are not seeing more litigation? Yes there are some notable cases but not a lot in the grand scheme of things. There are a few probable reasons for this:

First, the challenges involved in even knowing what has gone into training sets and which outputs have derived from copyrighted material. Without transparency and disclosure obligations how are labels, artists and others supposed to know that their rights have been infringed and how do they get access to platforms to interrogate them? This seems like the main blocker.

Secondly, rights holders may be adopting a ‘wait and see’ attitude and monitoring current legal cases, legislative developments, the enforcement of existing laws and the evolution of licensing deals.

Thirdly, commercially speaking, the big labels and artists may want their catalogues to be used provided that exploitation is accompanied by more licensing fees. They may not want to extinguish or kill off the use of their material altogether.

Finally, smaller independent rights holders may not have the financial ability to investigate and bring claims even if they could somehow get access to platforms to conduct tests.

Future legal claims

In England and Wales the basic statutory period for bringing a copyright claim is six years. So depending on how case law develops over the next few years there will still be a good amount of time for potential claimants to recoup losses going back in time unless, for example, those losses are wrapped up in new licensing deals.

Crystal ball gazing

There are people far more expert than I who may be able to predict what will happen in future. My best guess is that those with commercial leverage and financial muscle will get close to a win:win situation for both sides – platforms get the music to exploit and the rights holders get even more value from their catalogues via licensing fees as more and more people utilise music platforms. Hopefully smaller independents will be able to tuck into the ‘slipstream’ of these larger players and deals in order to receive their fair share. And above all perhaps, let’s hope that originality and creativity continues to flourish and is not deterred as the legal, technical and commercial aspects of all of this get straightened out.

Thanks for reading.

Michael Bywell, October 2026

Law of the Game: The Legalities of Game to Film Adaptations

Historically, adaptations have never been met with success. Whether it is a classic book turned feature film, or a feature film transformed into a TV show, there has always been tension and critical misfires when translating IP. Nonetheless there has been an exponential increase in popularity in gaming as a hobby; with innovative technology comes more sophisticated and complex design and plots that make these games so enticing. This, alongside the creation of accessible gaming platforms and devices, such as Steam, Nintendo Switch and PlayStation has led to a surge in videogame consumption. The film industry, consequently, has clearly seen how popular these gaming franchises are, in part due to their incredibly strong, loyal, and ingrained fanbases, and has found a new target audience to tap into. With these games already having an established community, studios can exploit this to lower the pressure faced when attempting to attract an audience to their work, which could make or break their opening weekend at the box office. And with a lot of games being more story-driven and narrative focused such as The Last of Us, Resident Evil and Fall Out, featuring lore, complex characters and plots, games would effortlessly synthesise with the film genre.

Though, in this new era of film, which seems to be gravitating towards the comfort of games, it poses the question of how much creative liberty the director has over the franchise, and how much control the creator of the game has over the film. Most people are overly focused on how faithful the adaptation will be to the plot, characters, and aesthetics of the game. As we have seen in the past, fans of established franchises are incredibly hard to please. If the casting for a character does not comply with the preconceived image that the masses have unanimously agreed on, the pressure on the studio increases in risk of losing their audience. There is a delicate balance between the fear of alienating old fans, risking a negative impact on the franchise, and making necessary changes for the game to translate accordingly into the film medium. Filmmakers must also keep in mind that not everyone watching their adaptation is part of the original fanbase and thus needs to translate the visuality and crux of the game into the film for newcomers. Adaptations benefit both IP and studios as we have seen through release of the films like Sonic, The Mario Movie, The Minecraft Movie, Mortal Kombat, Resident Evil and now Street Fighter which have allowed videogame companies to further explore not only their franchise, but the capacity of their IP and draw in new fans, which encourages more content to be created. Games are now the current focus of the adaptation movement, especially titles that are still considered active but how do we go from console to the big screen?

Firstly, what is IP? IP stands for intellectual property which is anything created such as inventions, art, literature, symbols, slogans, and designs which are protected by law. There are a few ways to protect your IP, such as copyright, trademarks, and patents. In this case, games usually encompass a mix of all these protective marks. So, when it comes to the adaptation of a game, which means sharing your IP, there is a process that enables both IP and studio to work harmoniously together. The process itself requires legal contemplation, as engaging with existing IP comes with the challenges of ownership. Producers must have the necessary rights to work with unoriginal media, which is where adaptation agreements come into play. An adaptation agreement is a foundational legal contract that grants a filmmaker the right to adapt IP-protected works into another film, TV show, or book. This contract embodies the duration of how long the rights last, the territory which the adaptation can be released and even what happens when a sequel or spin offs are pitched. The most recent example of a successful game to film adaptation, and the case study for this analysis is Zach Creggor’s Resident Evil, a highly anticipated film originating from its franchise of the same name.

To get from game to screen, there is a long history between the creators of the franchise and the live action rights. Capcom, founded in Osaka, Japan, in June 1983 by Kenzo Tsujimoto, is the company responsible for the creation of the RE games, and the owners of the entirety of the Resident Evil game franchise including its characters such as Leon Kennedy and Ada Wong, and the game’s entire universe and location such as The Umbrella Corp and Raccoon City. It was only in 1996 when Don Friedman, a licensing coordinator for Capcom, contacted Marc Mostman (ELA) to pitch the game as a film, seeing its cinematic protentional from beyond a console. In spite of having a strong pitch and marketing strategy, every major studio in Hollywood rejected the proposed project. This was until the German production company Constantin Film heard about the idea and accepted it immediately. This partnership catalysed the 2002 release of the first Resident Evil movie directed by Paul W.S. Anderson. Under their adaptation agreement, it was established that Capcom still owned the rights of the game, while Constantin Films owned the rights to the live action universe of the franchise. The agreement also made sure that the studio agreed to make and release live action films, or the rights would revert to Capcom. A major clause in the contract underlined by Capcom is that Constantin films cannot make a live action movie involving the games major protagonists/antagonists or any primary game story arcs, and instead can make parallel stories or separate universes that connect with the main narrative. This is why there has yet to be any live action films involving the franchises favourite characters.

Much like the adaptation agreement between Capcom and Constantin Films, both companies have a different share in merchandise. While Capcom is still the legal owner of the original IP, thus controlling the official cooperate merchandise, Constantin Films take charge of the film’s specific merchandise rights that are directly tied to the live action universe of the Resident Evil franchise; this includes ‘medic bag’ popcorn buckets that mirror the ones used by characters in the film. Constantin Films also has a long-standing relationship with Sony Productions as a distributor, and for this specific 2026 rendition of Resident Evil it marks the first time PlayStation has been involved in the franchise as the game has historically been known for being played on Sony devices.

In most cases when creative approval rights are discussed, the game developers often discuss scripts and characters beforehand to avoid tarnishing their image, but for the newest addition to the franchise, Zach Creggor (director of Barbarian and Weapons) was given full creative control by Capcom themselves. He avoided using characters already established in the franchise, and despite having said in interviews that he was nervous to take on such a feat in case he would upset the fans, the film was an overall success and became the biggest horror film of the year. Creggor placed careful odes to the original game franchise with the first-person shooter perspectives, and the dreary atmosphere. Despite not utilising any of the games original plots and characters, Creggor made the film feel a part of the franchise, involving zombie like mutations, a cure, and the infamous pharmaceutical and biotechnology conglomerate from the game (Umbrella Corp) which tied in the franchise to this new addition to the live action universe.

Game to film adaptations are overall a great way to expand the universe of existing IP beyond the console and allows for a different creative perspective to give new life and meaning to established worlds, and with the success of the collaborative efforts on the Resident Evil film I hope to see many more IPs branch out into new mediums.

Fabiola Stannard-Parages, October 2026

Behind the Eyes: Smart Glasses, Data Protection, and The Step Back

With their neutral hardware, hidden camera, and strategic partnerships with the fashionable labels Ray-Ban and Oakley, Meta glasses may appear to the untrained eye as just another pair of lenses. In fact, the lenses have been the source of considerable controversy – frequently referred to as ‘pervert glasses’. The people of London have made their stance on the matter clear, placing fake, satirical ‘Smart Glasses Amnesty’ signs on public bins, creating stickers that mimic official TfL signage aimed at the wearers of Meta glasses, and downloading the app "ZuckOff" on mass, which warns app-users of nearby smart glasses. 

Smart glasses are not illegal. In fact, the specs have had a considerable positive impact on many. They have been revolutionary for the blind community – the glasses can read out visible text from one's surroundings and describe in detail what is going on in the space around the wearer. Aware of the positive impact they were having, Meta donated 15,000 pairs of glasses to a charity for the blind in Ireland. The blind community is not the only community excited by the concept of the glasses – many professions will benefit too. In the travel and hospitality sector, for example, the glasses will enable real-time language translation, and tradesmen and beauticians may benefit from being able to record and share their work as they do it. But it is exactly this – the ability to record what is happening around you and then instantly upload it – that has triggered the negative response, sparking conversations around consent, misuse of information, data protection and legality. 

To quote tech founder Liz Hunter, the glasses ‘have rolled back [women's] privacy rights right before our very eyes’. Indeed, a pattern has emerged in which women have been filmed by men wearing smart glasses without their consent, before the footage is then posted on social media. Often the women targeted are unaware that they have been filmed until they find the video posted online. This covert filming has taken place in both public and private settings. In a recent case at Warrington Magistrates' Court, David Williams pleaded guilty to voyeurism after using smart glasses to record a private sexual encounter without the victim's consent.  

Data Protection and the UK GDPR 

It is one thing for the glasses to capture content; it is another for them to process the information they gather from accompanying their owner around during the days, weeks and private moments. For those who wear their smart glasses to a space which makes their special category data known – health, religion, political opinions, trade union membership, racial or ethnic origin, sex life or sexual orientation – it is vital to understand what is being done with that information. It is also important not to underestimate the ability to establish special category data from daily wear. When the glasses record your journeys, the people you spend your time with, your weekdays and your weekends, the complex system will likely make inferences about special category data. 

Data protection laws regulate how this kind of data is used. The UK GDPR applies where its scope is engaged, which it will be unless an activity is purely personal or household. However, if private surveillance includes public information the UK GDPR will apply (C-212/13 Ryneš v Úřad pro ochranu osobních údajů). Therefore, the UK GDPR will apply if the wearer of the smart glasses is using them for work, or to record strangers. If the UK GDPR is engaged, the controller must identify a lawful basis, act fairly and transparently, minimise data, limit purpose and retention, keep data secure, respect rights, and be able to demonstrate compliance. 

The easiest way to ensure that the wearers of smart glasses act fairly and transparently is to ensure that the LED recording light is activated at all times. Although Meta have attempted to make it considerably more difficult to disable the recording light, 1% of wearers have been able to do so in the hope of stopping the recording light from showing. Other wearers have obstructed the light altogether. This appears to be a problem that Meta have acknowledged, as on the 24th of September Meta launched an audio-only set of lenses with no camera, which many have seen as a response to the concerns around data protection that have stemmed from the previous lenses. 

In an attempt to minimise the data collected, the ICO have encouraged conversation around why recording is needed. Although there may be some uses of the glasses that are necessary in certain industries, it is important to assess whether less invasive methods can be used to achieve the same outcome. Similarly, it is vital to identify a lawful basis if special category data is collected. Continued recording – even for professional purposes – is unlikely to be lawful. 

Increase in crime 

The glasses have created fear around an increase in crimes like voyeurism, upskirting and intimate recording. As this article has clarified, smart glasses are legal. However, they have been used in illegal ways. In response to a rise in cases in these areas, the intimate image regime has developed – in February 2026, section 138 of the Data (Use and Access) Act 2025 was amended to ensure that the creation of intimate images of adults was a punishable offence. 

Conclusion 

It is undeniable that technology is developing alongside the law. Statutory provisions governing data protection laws will have to keep developing, and it is safe to say that in the coming days, weeks and months, the landscape will be different. The recent introduction of glasses without a camera demonstrates that tech companies too are in unknown territory – is this a strategic step away from the model that caused an uproar, or is it a continuation of a product that is the first of many? 

Dolly Payne, September 2026

The Digital Double and the Public Domain: Who Controls a Performer’s Image After Death?

Picture watching a new film and seeing a familiar actor on screen - except the actor had died years ago. They had never filmed the scene, and had never agreed to appear in it. This is no longer a particularly hypothetical scenario.

Generative artificial intelligence (“AI”) is now capable of producing a performance from a deceased actor. By reproducing an individual’s distinctive features, these systems can place performers in works in which they never participated. Such posthumous digital performances stand apart from contemporaneous projects created by the performer during the course of their lifetime. A digital counterpart will attribute an expression based on predictive patterns that the person depicted may neither have contemplated nor authorised. As these tools become more accessible, the legal and ethical framework required to regulate their production and commercial exploitation becomes correspondingly uncertain.

The central issue, therefore, extends beyond protecting deceased performers from isolated “deepfakes”. It concerns whether the law should recognise a distinct posthumous interest in digital identity and, if so, who can allocate that interest without turning personality into a perpetual form of intellectual property (IP) or unduly restricting expression in the public domain.

When the Performer Lives on as a Digital Double

A performance is not the same thing as a person; UK law grants performers a series of statutory rights designed to facilitate their commercial exploitation. Under the Copyright, Designs and Patents Act 1988 (CDPA) [1], those rights may be infringed by the unauthorized recording of the whole or a substantial part of a qualifying performance (s.182), including public dealings involving an unauthorised recording (ss.183-184). These rights exist independently from copyright in the underlying work.

What these provisions do not establish, however, is a general proprietary right over the performer’s identity. The distinction matters because appropriating the value associated with a performance is not necessarily the same as copying the performance itself. Copyright and performers’ rights may be capable of addressing the latter, but an AI replica can raise a different question: whether the law should protect the commercial and expressive value embodied in the performer’s identity even where no recognisable part of an original performance has been reproduced.

AI brings this performative distinction into sharp focus. Where a system reproduces an existing recording, copyright or performers’ rights may provide a basis for challenging its use. The position is less straightforward where the system generates an entirely new performance that merely looks or sounds like a deceased actor.

The UK Government’s 2026 report acknowledges this gap, noting that performers’ economic and moral rights under Part 2 CDPA attach only to recordings of existing performances [2]. It may not assist where AI generates a new performance containing a digital replica of the performer. The Government is therefore considering whether additional protection for digital replicas or personality is necessary. After all, the Rome Convention [3] protects performers and their fixation rather than a general right over the performer’s identity.

The result is a strange legal distinction: the law can protect what the performer actually did, without necessarily protecting the person who did it.

What survives when the performer is no longer here to object?

The problem surrounding this distinction only becomes more complex after death. The 2024 dispute concerning George Carlin provides a useful example [4]. Carlin’s estate sued the creators of an AI-generated comedy routine which imitated Carlin’s voice, style and opinions, alleging both copyright infringement and violation of his publicity rights. The litigation settled, with the defendants agreeing to remove the material and refrain from using Carlin’s likeness. The settlement therefore established no judicial test governing AI-generated replicas. Nevertheless, the dispute demonstrates how distinct legal interests may converge where an AI output draws simultaneously upon a deceased performer’s existing works and the identity associated with them [4].

That leaves a more fundamental question. Is the law required to give the estate control over the performer’s identity at all?

American publicity-rights law shows how difficult the answer can be. In Memphis Development Foundation v Factors Etc Inc [5], the Sixth Circuit held that Elvis Presley’s right of publicity did not survive death and that the ability to commercially exploit his identity therefore passed into the public domain. Other US courts reached the opposite conclusion, most notably in Elvis Presley International Memorial Foundation v Crowell [6], in which Tennessee law recognised a posthumous right of publicity.

The contrast is revealing. A performer’s identity can be understood either as an interest which ends with the individual or as a commercial property right capable of surviving them.

Shaw Family Archives Ltd v CMG Worldwide Inc [7] exposes another difficulty: inheritance cannot simply be assumed. In the Marilyn Monroe litigation, the court held that Monroe could not have transferred by will a property right she did not possess when she died in 1962, when neither New York, California nor Indiana recognised a descendible postmortem publicity right.

These cases expose the weakness of relying upon an unrestricted property model. If a persona is treated as property, difficult questions immediately arise over who owns it, whether it can be inherited, how long it lasts and whether subsequent legislation can retrospectively create rights in an identity that did not exist at death. For AI, those questions are no longer theoretical.

The Public Domain Problem

There is a strong argument for resisting perpetual control. A digital replica can make a deceased performer appear in something they had never encountered, even if that extends to endorsing a product or expressing misaligned political opinions [8]. Access to this technology can do more than reproduce an image: in some sense, it can manufacture a new version of a person.

Posthumous reproduction can alter the relationship between a performer and their public perception. Considering Mark Bartholomew’s ‘A Right to Be Left Dead’ [9], this concern is met with legitimate reservations. It is argued that postmortem protection for digital reanimation cannot simply replicate the legal treatment of living individuals. On balance, a proposal for a narrowly constructed right that includes a minimal term and a requirement of prior exploitation, would aim to preserve artistic innovation and technological development.

After all, the public domain is not simply a place where rights are forgotten. It is what remains available for other people to use, reinterpret and build upon. A perpetual right over a celebrity’s image could therefore transform personality into something resembling an indefinite intellectual property right.

US publicity-rights cases provide some useful ways of thinking about that distinction. Rogers v Grimaldi provides one established way of thinking about that boundary. The Second Circuit protected the use of a person’s name in an expressive work where it had ‘artistic relevance’ [10] and was not explicitly misleading as to source or endorsement. Similarly, Comedy III Productions v Gary Saderup developed California’s transformative-use approach, asking whether a work containing a celebrity’s likeness has been sufficiently transformed into the creator’s own expression rather than remaining principally an appropriation of the celebrity’s likeness. The principle is valuable for AI because a digital recreation may be used to communicate something about a performer rather than to exploit the performer commercially [11].

The opposite extreme is illustrated by Zacchini v Scripps-Howard. There, the Supreme Court demonstrated that appropriating the entirety of a performer’s act could still be actionable notwithstanding its presentation as news [13].

The lesson for AI is not that one side should always win: it is that purpose and context matter. The law should be particularly sceptical of a digital replica whose principal purpose is to appropriate the economic value of the performer, while remaining cautious about restricting works that use the replica as part of new expression.

Towards a Limited Right of Control

US legislative responses increasingly treat a digital replica used to sell a product very differently from one used in works such as a documentary. The proposed NO FAKES Act [14] would create a federal property right in highly realistic digital replicas, while providing exceptions for uses including news, documentary, commentary, criticism, satire and parody [15]. Across the jurisdictional pond, the UK could respond by introducing a specific statutory right that need not require absolute control over expression [16].

The UK now has an opportunity to answer that question before the technology accelerates beyond a point of regulation. A new right could be justified, but it should not become a general property right in personality. Instead, legislation could focus on the clearest form of misappropriation: the unauthorised commercial use of a recognisable digital replica which exploits the performer’s identity or falsely suggests their endorsement or participation. In doing so, the protection should also leave room for legitimate expression - including journalism, documentary, criticism, parody and historical or artistic uses. The US experience indicates that an effective regime will need some operative mechanism to balance publicity interests against freedom of expression.

Development of a limited posthumous term would recognise that estates may have legitimate interests in managing a performer’s legacy, while acknowledging that those interests cannot automatically outweigh the public’s eventual ability to engage with cultural figures freely. Bartholomew's proposal for a structured postmortem right provides one such possible model. This would also regulate a particular kind of “appropriation of identity” which existing copyright and performers’ rights may not capture.

The Government’s current approach points towards precisely this debate. The March 2026 report acknowledges that existing UK law leaves gaps and is considering whether new protection for digital replicas or personality is appropriate, while recognising the need to preserve legitimate innovation. Ultimately, achieving the appropriate balance is the central challenge, although it is achievable. The law should aim to protect individuals from having their identity commercially appropriated, without making identity itself an endless form of IP.

The digital double may allow a performer to appear long after death, but that does not mean their legal identity should live forever under someone else’s control. Eventually, even a digital double may enter into the public domain.

Harvey Read, September 2026

1: Copyright, Designs and Patents Act 1988

2: https://www.gov.uk/government/publications/report-and-impact-assessment-on-copyright-and-artificial-intelligence/report-on-copyright-and-artificial-intelligence

3: https://www.wipo.int/en/web/treaties/ip/rome/summary_rome

4: https://www.reuters.com/legal/transactional/george-carlins-estate-settles-lawsuit-over-ai-generated-comedy-routine-2024-04-03/

5: https://law.justia.com/cases/federal/district-courts/FSupp/441/1323/1427639/

6: https://eprints.whiterose.ac.uk/id/eprint/180330/1/eslj-708-wall.pdf

7: https://law.justia.com/cases/federal/district-courts/new-york/nysdce/1:2005cv03939/310990/106/

8: https://www.tate.org.uk/research/tate-papers/08/posthumous-legal-and-ethical-issues

9: https://www.californialawreview.org/print/left-dead

10: https://www.skadden.com/insights/publications/2023/06/supreme-court-sharply-limits-applicability

11: https://wfujournaloflawandpolicy.org/deepfakes-of-the-dead-applying-postmortem-publicity-law-to-artificial-intelligence-digital-replicas/

12: cf.11

13: https://scholarlycommons.law.hofstra.edu/cgi/viewcontent.cgi?article=1245&context=acteclj

14: https://www.reedsmith.com/articles/ai-and-publicity-rights-the-no-fakes-act-strikes-a-chord/

15: https://ssrn.com/abstract=4635039

16: https://www.vwv.co.uk/insights/articles/copyright-and-artificial-intelligence-the-uks-pivot-towards-a-new-digital-replica-right/

AI Training on Copyrighted Content: Should other Nations’ Courts learn from Germany?

AI has increasingly been used to generate music and entertainment content, but how are applications like Suno and Udio able to generate ‘new’ works? As former MP, Kevin Brennan, explained, AI learns ‘the patters of human creativity ... based on algorithms and predictions’, in order to generate new content. The problem this poses for the entertainment and music industries is that without licenses, AI companies would be exploiting and damaging the industries’ very foundation: Intellectual Property (IP).  

Copyright law specifically, serves as an integral component of a functioning creative landscape. From songwriters to film and tv producers to authors and more, rightsholders hold the power to exploit their IP and capitalise on it. When a legal person uses copyrighted work without the authorisation of all the rightsholders, there may be copyright infringement. In the context of AI training on copyrighted works, copyright infringement is of major concern globally.  

There is significant legal tension between AI companies and record labels, production companies, artists and other rightsholders in the entertainment and music industries.  There have been a number of lawsuits and licensing deals recently, however the Munich Regional Court’s ruling at the end of July 2026, marks an important landmark for the problem of AI training on unlicensed works.  

The Facts of GEMA v Suno Inc:  

GEMA, the German music collecting society, brought an action against Suno, who provides an AI music generator, in January 2025, after Suno failed to respond to GEMA’s request for licensing. The case concerned six compositions which GEMA had in its repertoire: “Daddy Cool”, “Mambo No. 5”, “Forever Young”, “Atemlos dur die Nacht”, “Big in Japan” and “Rasputin”. In bringing the lawsuit, GEMA sought an injunction against four copyright rights-related uses of the works.  

In the US, Suno trained its AI models on copyrighted works, including those in dispute. The works were obtained by stream-ripping on YouTube, which involves going around YouTube’s downloading restriction. This means that Suno did not acquire licensed music to train its AI, and this was not in dispute in the case. After obtaining unlicensed musical works, Suno attached metadata to those audio files and converted them to numerical units which Suno’s AI model’s parameters trained on. However, the final model, which is stored in Germany, does not itself contain the original audio files.  

The issue in this case was not establishing if Suno had in fact trained on copyrighted works without a license, rather, it was that Suno argued that it did not need a license in the first place. Suno challenged the jurisdiction of the German court because the training was done in the US. Additionally, Suno tried to rely on both US fair use defence and the EU’s text and data mining exception. The ruling in this case is therefore of high importance in the current increasingly technological landscape.  

The Munich District Court’s Ruling:  

Firstly, it should be noted that the German court was able to rule on the training of the AI model even though it happened in the US, because of provisions in the German Collecting Societies Act, s 131(1)-(2), and the German Code of Civil Procedure, s 32. There are restrictions to this as only collecting societies may benefit from this law and there must be an infringement that took place in Germany.  

The court found that there was reproduction of works because the AI model memorised them and exact copies were not needed. By inputting a style with lyrics and a title and comparing the AI outputs to the original works that were the subject of the case, memorisation was proven. Notably, Suno argued that the AI-generated infringing outputs were the user, GEMA’s, responsibility because the prompts that were given to the AI were “complex”. The court rejected this. The prompts were not close-ended and the AI model was free to decide important aspects of the final output: the harmony, tempo, melody and rhythm. Using the reasoning of the GEMA v OpenAI case that the same court ruled on previously, the works were fixed in numerical probabilities and exemplified in parameters that were specified, and so they were reproduced.  

The court determined that there was no text and data mining exception in this case. The reproductions were not for data analysis and Suno, by going around YouTube’s downloading restriction, had obtained the works unlawfully. Thus, no text and data mining exception could be relied on by Suno.  

The German court also ruled on whether the training of Suno’s AI model in the US infringed US copyright law. In doing so, reproduction was again established and the court analysed if the four fair use factors (the work’s nature, the purpose but also character of the use, the amount of the works used, and the market impact (17 U.S.C. § 107)) could be weighed against Suno. Importantly, the court distinguished the facts of GEMA v Suno, with Bartz v Anthropic, because in Bartz the AI did not output its training data. The court found that none of the four factors aided Suno in proving that they did not need a license. Their use of the works was commercial not transformative and there was bad faith in acquiring the music unlawfully. The compositions were fully copied and Suno had taken the expression of the works.  

Why is this Important? Should the Courts of other Nations also hold that Training AI Requires Licenses?  

GEMA v Suno is a landmark victory for creatives and other Nations should consider requiring AI companies to obtain licenses to train their models on copyrighted content. The ruling on US fair use does raise the question: what would a US court have thought? If some record labels like Warner Music Group did not settle in their infringement suits against Udio (another AI music generator) and Suno, with licensing deals, a US court may have applied the same analysis as the German court in GEMA v Suno. Fortunately, there is still a RIAA (Recording Industry Association of America) coordinated lawsuit against Sony Music and Universal Music Group and other litigations in the US. This means US courts can adopt the German court's reasoning, and if they do, those across the music, entertainment and publishing sectors that have licensed may have chosen to do so too soon.  

Additionally, it is hopeful that European nations also find infringement and that a text and data mining exception is not available to AI companies that unlawfully obtain copyrighted material to train their AI models. While there is hope for creatives to protect their work from unlicensed AI training, the factual nuances of GEMA v Suno, namely that the case focused on only six compositions and there was proof that Suno could reproduce those exact compositions, should be noted. Where future cases have a significantly larger volume of copyrighted works in question, it will be harder to evidence reproductions of all works in question. Nevertheless, the German case provides optimism for the protection of creative works, which is crucial for the entertainment and music industries as IP is their foundation.  

Sejal Patel, August 2026

Sources:  

APPG on Music Report: ‘Artificial Intelligence and the Music Industry – Master or Servant?’ https://www.ukmusic.org/wp-content/uploads/2024/04/APPG-AI-Report-Low-res.pdf  

https://www.twobirds.com/en/insights/2026/germany/munich-district-court-rules-on-ai-generated-music-gema-v-suno  

https://www.simmons-simmons.com/en/publications/cmszvaaz6003oug849s2sqawl/gema-v-suno-ai-key-eu-decision-on-ai-training-outputs-and-copyright  

https://www.forbes.com/sites/virginieberger/2026/08/05/suno-lost-to-gema-why-the-ruling-should-worry-ai-music-companies/  

https://www.musicbusinessworldwide.com/suno-infringed-copyright-in-gema-case-german-court-rules/  

AI – The Secret Author: Would You Want to Know?

Your favourite magazine – whether it is Vogue, National Geographic or another trusted publication – may already contain AI-generated content without notifying you. As Artificial Intelligence becomes embedded in newsrooms, an important question emerges: should readers have a right to know?

81% of publishers now use Artificial Intelligence for production or in editorial and almost 1 in every 5 publishers use Artificial Intelligence to assist with personalisation in their published content (1). Artificial Intelligence has become a common feature in modern publishing ranging from assistance with spell checks to generating whole pages of content. Therefore, would this revelation impact your trust with that magazine, or would it make little difference?

As we continue to read and spread information drawn from published content, the impact of Artificial Intelligence in this area cannot be understated. The ongoing issue of hallucinations in AI-generated content poses a threat to journalistic integrity and requires keen oversight by journalists to prevent the spread of false information.

The Role of AI

Artificial Intelligence can be used to assist journalists with their first drafts, generate snappy headlines, grammar check their work, find key takeaways, create images and personalise content. Applications are also adopting Artificial Intelligence such as BBC iPlayer: “The evolution of AI in all its forms offers tremendous opportunities for creativity, innovation and improved productivity in every area of the BBC” (3). While concern grows for the future of jobs in journalism, Artificial Intelligence may only play a minor role in assisting journalists to produce higher quality work in a more efficient fashion.

Benefits of AI in Journalism

There are numerous benefits to using Artificial Intelligence in journalism, such as a faster production rate, cost-effectiveness, and the ability for journalists to stay on top of active stories. However, 62% of UK journalists view Artificial Intelligence as a large threat to the future of journalism (4).

On the other hand, Vogue’s August 2025 issue has caused severe backlash and criticism due to a discreet label stating, “Produced by Seraphinne Vallora on AI.” The use of Artificial Intelligence in such a long-standing and esteemed publication has caused an “outrage” (2) amongst readers, with many “long time subscribers announcing cancellations” (2). The most prominent reason for this extreme backlash is the desire for authenticity.

The negativity following Vogue’s use of Artificial Intelligence is an example of why publications are more likely to shy away from publishing AI-generated content. Even with the potential of backlash, should we expect publishers to notify their readers whenever Artificial Intelligence has been used within their publications?

Transparency

Journalism relies on journalistic integrity as a foundation to keep audiences engaged. Disclosing when Artificial Intelligence has been used helps avoid deception and enables readers to research and assess the credibility of content. As Warren Buffett noted, “It takes 20 years to build a reputation and 5 minutes to ruin it” (5); therefore, disclosure may be a necessary step to preserve a publication’s reputation.

However, should we require transparency where Artificial Intelligence has been used merely as an editing tool? Grammarly, Photoshop and other known editing tools do not require a label to expose usage - should Artificial Intelligence require such disclosure?

Legal Complications

The Law is yet to catch up with the rapid development of Artificial Intelligence in journalism, creating further legal complications:

• European Union Artificial Intelligence Act 2024 - Article 50 “Deployers of an AI system that generates or manipulates text which is published with the purpose of informing the public on matters of public interest shall disclose that the text has been artificially generated or manipulated.” (6). Under the new legislation, it is mandatory to label AI-generated content such as deepfakes; however, the disclosure of AI being used as an editing tool or for general writing assistance is not mandated under this Act.

• Copyrights, Designs and Patent Act 1988 – Section 9(3) states that “In the case of a literary, dramatic, musical or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken” (7). However, the increasing sophistication of generative AI has prompted debate about whether this provision remains appropriate. How can we really give journalists full praise for authorship of “their” work if it was mostly computer-generated?

It is unlikely that Artificial Intelligence will disappear from journalism, especially for simple editing purposes. However, the need for clear and transparent labelling may be the key step in ensuring readers can continue to place trust in publications and potentially prevent some of the backlash seen with Vogue. If the core of journalism is trust and integrity, disclosure of AI-generated content should be seen as a necessity, and not just a source of anxiety for journalists.

Hamsini Bacchu, August 2026

1: https://presenc.ai/research/ai-in-media-and-publishing-statistics-2026

2: https://www.forbes.com/sites/moinroberts-islam/2025/07/29/vogue-erupts-ai-generated-models-spark-reader-fury-and-industry-panic/

3: https://www.bbc.co.uk/editorialguidelines/guidance/use-of-artificial-intelligence#editorialissuesintheuseofai

4: https://reutersinstitute.politics.ox.ac.uk/news/speed-hoaxes-and-mistrust-how-ai-transforming-freelance-journalism

5: https://www.forbes.com/sites/erikaandersen/2013/12/02/23-quotes-from-warren-buffett-on-life-and-generosity/

6: https://artificialintelligenceact.eu/article/50/

7: https://www.legislation.gov.uk/ukpga/1988/48/section/9

Behind the Scenes: the AI Revolution in Film & TV

Imagine writing a script and seeing it storyboarded by a computer in seconds, or watching it create an animation instantly; it is the modern reality of filmmaking today. Artificial intelligence is no longer just a futuristic, sci-fi device – it is actively reshaping our lives as we speak. This includes the making of our favourite films and TV shows. But as technology continues its rapid growth, the industry faces a critical question: will AI make or break the soul of cinema?

AI is already being deployed in some areas of the film and TV production process, including the conceptualisation of new ideas, storyboarding, and visualisation. One may argue that utilising advanced technology such as AI is rather beneficial to the filmmaking industry, reducing time and human labour whilst producing first drafts and pre-production results. However, studies have shown that uncertainty arises when considering the extent of AI and how it will change production, extending to how those changes manifest throughout the content and distribution ecosystem. As a result, industry leaders face the potential threat of AI materially altering the industry’s structure and profit pools.

McKinsey & Company’s Alec Wrubel provided an insight into the perceived benefits of AI usage (1). It is commonly believed that with the assistance of AI and similar tools, better quality results are produced at an accelerated rate and lower cost. Wrubel draws on the film Star Wars: A New Hope, in the 1970s, in comparison to The Rise of Skywalker (2019) to challenge this assumption. He states that whilst the introduction of CGI resulted in increased visual fidelity, they did not reduce the overall cost of filmmaking. Instead, studios often reinvest these technological efficiencies into creating more visually ambitious films. As a result, rather than replacing expenditure, incorporating technology like AI can be argued to enhance creative output rather than simply reducing production costs. Then, how useful is AI, really? Do the benefits outweigh the flaws?

AI’s continuous advancements result in its increased accessibility, extending far beyond its presence in the filmmaking industry. The prevalence of AI can also be seen in open platforms like YouTube and TikTok, where a variety of AI-generated content is viewed every day. This is where deeper issues can potentially arise, conflicting with legal aspects such as intellectual property (IP), liability, and consent. From a critical perspective, AI can be perceived as a “threat”, when considering deepfakes and AI-generated performers being able to replicate micro-expressions and vocal nuances effortlessly, without human involvement (2). Applications such as CapCut are widely available to the public, granting them access to the effortless creation of AI-generated videos. In comparison to those in the filmmaking industry, a sense of professionalism and legality is lost in the production process. As a result, it remains ambiguous who is liable when legal conflict arises.

That said, it is not wrong to address the extent of creativity AI can provide. As Jamie Vickers claims, “almost every major technological innovation in media has produced a format that nobody envisioned”. Essentially, even TV itself is a major new art form that we all have grown to love. Then there was the introduction of social media platforms, such as Instagram, allowing for free, creative self-expression. From this perspective, it is arguable that AI isn’t a replica of human creativity; it is a tool that can expand authentic human creativity and increase the value of content creation and storytelling. Considering this view, it is crucial to note the importance of balance. The use of AI poses as many benefits as it does threats; the key is to utilise it sparingly, avoiding ethical misuse (3).

One of the major concerns regarding AI is its impact on employment in the film and TV industry. Writers, actors, animators, editors and visual effects artists have all questioned whether AI could eventually replace aspects of their career. While AI has the potential of completing repetitive tasks successfully, many creatives argue that the storytelling of films is predominantly reliant on unique human experience – qualities that AI cannot replicate. Rather than replacing filmmakers, AI can be used collaboratively. Efficiency can improve in initial processes like generating first drafts and ideas, whilst simultaneously retaining authentic human creation. Therefore, the success of AI in the industry is largely dependent on humans’ responsibility towards using advanced technology and its implementation.

Ultimately, AI is neither the saviour nor the downfall of film and television. It offers significant opportunities to improve filmmaking processes, inspire creativity, and make production more accessible. However, it can also present ethical and legal challenges surrounding issues such as consent, authenticity, intellectual property, and liability. As AI continues to evolve, it is inevitable that it will become deeper embedded into our lives. The future of cinema will rely on maintaining a healthy balance between technological innovation and human creativity. If used ethically and with appropriate regulation, AI has the potential to enhance the future of filmmaking without replacing the people whose unique imagination gives cinema its soul.

Crystal Kan, July 2026

1: https://www.mckinsey.com/featured-insights/mckinsey-explainers/lights-camera-algorithm-how-ai-is-rewriting-the-rules-of-film-and-tv

2: https://ipwatchdog.com/2025/04/21/ai-ip-hollywood-finding-balance-verge-new-creative-class/

3: https://medium.com/@jeetpadhya35/ai-is-a-tool-not-a-replacement-how-artificial-intelligence-amplifies-human-capability-7a798ba864d2

Music Copyright and AI: What Can We Learn from the Amen Break?

It is highly likely that in the past half decade, you have unknowingly listened to music made partially, even entirely, by AI. Whether it’s a producer using AI instruments and vocals (1) or completely AI-generated artists (2), music is rapidly beginning to include more sounds created on AI driven software.

With listeners expressing their concerns when it comes to identifying AI usage in their music, a similar issue emerges regarding copyright. For example, in what way is using recordings to train music-generating models copyright infringement? Or how should the use of such software be made clear as well as credited? However, what if these seemingly modern concerns didn’t emerge as recently as one would’ve thought.

Going back 70 years to 1956, long before any Zoomer used ChatGPT on an iPad, chemistry professor Lejaren Hiller was busy tinkering with the Illiac I, the University of Illinois Urbana-Champaign’s only computer at the time. Instead of cracking codes or opening search engines, Hiller had programmed the computer to generate its own musical compositions, using what is now known as algorithmic composition. The piece String Quartet No 4, originally titled the Illiac Suite, was the first substantial piece of music to be composed not only on, but by a computer. It was performed by four students, one of whom remembers a packed auditorium of people “who showed up to see what the monster of a computer could do” (3), with one of Heller’s former students noting that he had “touched a nerve in a very deep way”.

Even then, at the very first instance of something that resembled technology to come, the reaction was one of apprehension. What does this mean for the future of music? Would computers eventually take over the arts? At what point won’t we be able to tell the difference between what a human and a computer makes (4)? Before that point however, computers were busy chopping up and changing old sounds and making them into new ones.

In 1969, Gregory Coleman played a 7 second drum break that would eventually be the most heard drum sound in the world. The track he played on, ‘Amen, Brother’ by American soul group The Winstons, has as of today been sampled on over 7000 tracks, making it the undisputed most sampled track in history (5). From hip-hop to jungle, drum and bass to rave music, a wide span of genres utilised the drum beat and made it their own, such as a rap loop on Salt-N-Pepa’s ‘I Desire’ and NWA’s ‘Straight Outta Compton’ or chopped up on jungle tracks like Lennie De Ice’s ‘We are I.E.’ (6).

The beat, coined the ‘Amen Break’ most notably re-emerged on a 1986 compilation named Ultimate Beats and Breaks, while also appearing on other sampling collections. With the 80s boom of hip-hop and electronic music, record companies created these compilations including songs with beats that were good to scratch and mix. These companies, however, were effectively selling the ‘Amen Break’ as their own copyrighted material, and by the 2000s multiple copyrights existed for the track. The lack of strict copyright laws at the time meant that any original artist was unlikely to see a reward for their work. On the other side of the coin, this flexibility enabled the ‘Amen Break’ to inspire thousands of minutes of music and new genres, far beyond those original 7 seconds.

Hence, while the free use of the ‘Amen Break’ would contribute to the development of multiple music scenes, Coleman would see no royalties or reward in his lifetime, dying homeless and penniless on February 5th, 2006. In 2015, a crowdfund raised £24,000 to give to The Winstons’ frontman Richard Spencer, as a rightful reward for their track’s success (7).

In the modern day, have we not returned to a similar issue? Just as music copyright law didn’t keep up for the likes of Coleman in the age of sampling, we’ve seen the law once again lag behind the booming growth of AI music software. However, could intense scepticism around AI as well as copyright crackdowns lead to regulatory control that inhibits potential creativity?

Both Suno and Udio, the two most popular AI music generators, have been sued by Universal Music Group (UMG) and Sony Music Entertainment amongst multiple other entities in the past 2 years for copyright infringement. The original claims state that their AI models had been trained on copyrighted music without authorization, which after 2 years of discovery has led to both UMG and Sony adding over 60,000 recordings to their lawsuits against Suno (8). Furthermore, despite having an initial bid denied, Sony have once again sued Udio asserting over 30,000 recordings have been illegally copied (9). The result of these cases is likely to shape the future for creative AI models, deciding whether training these models on data sets including copyrighted recordings without license is deemed fair use or not.

In opposition to the labels, the AI companies point to recent rulings such as Bartz v. Anthropic, where the use of books for AI training was ruled as fair use, as well as Kadrey v. Meta Platforms where a similar ruling was found. In the Anthropic case, the judge deemed that the AI’s output was not similar enough to the source books, calling the use “spectacularly transformative” (8). However, in the cases with Suno and Udio, the argument remains as to what extent the music created is too similar to the recordings used to train them.

There may be answers arriving soon, especially as previous claimant Warner Music Group (WMG) settled and launched a joint venture with Suno to create data sets including licensed tracks for their AI model (10). UMG have similarly started to settle with Suno and are in the process of licensing tracks for their data sets also. Perhaps then, a suitable compromise will be met that enables these AI music programs to operate legally. However, if artists do not approve of their music being used for these models, they’ll be left with less recordings to train from and will likely be less effective tools. But if doing so protects these songs from being copied an infringing way, then it is just as important to protect artists and their work – protected so that the only time their work is copied, they’ll be rewarded accordingly.

For those artists who are using these AI programs as either inspiration for their work (11) or as tools for their creative process (12), a future where these programs are both law and label approved could be a positive for enabling more streamlined creativity for musicians wanting to use these tools. Although, a further issue then emerges regarding transparency about AI use, with listeners wanting to know the extent to which their music is AI generated (13). Then with copyright, if artists do use these AI music-making platforms, to what extent is their music their own, or is owned by the companies behind these softwares.

With this, perhaps we should think back to the ‘Amen Break’, how it led to the creation of new songs and genres inspiring countless musicians, but also how its original creators saw none of the success land in their own hands. But what if the story wasn’t so simple. What if Coleman didn’t even own the drum break he played. The song ‘Amen, Brother’, is in fact an up-tempo adaptation of ‘Amen’, a gospel tune by Jester Harrison (14). In this case, if every song is inspired, copied, or updated even before these AI programs existed, then is the process of transforming the old into the new not just the process of human creativity? What is being judged then, is truly how similar the AI process of creativity is to the human one. On the one hand, we must encourage this creativity and allow for new music to be made. On the other, we must protect the music and artists that have already been.

Theo Grange, July 2026

1: https://aristake.com/ai-tools-musicians-study/

2: https://www.rollingstone.com/music/music-features/timbaland-new-artist-tata-ai-1235356185/

3: https://www.theguardian.com/music/2021/dec/07/he-touched-a-nerve-how-the-first-piece-of-ai-music-was-born-in-1956

4: https://newsroom-deezer.com/2025/11/deezer-ipsos-survey-ai-music/

5: https://www.whosampled.com/The-Winstons/Amen,-Brother/

6: https://www.youtube.com/watch?v=wusSmIV-FE8 https://www.youtube.com/watch?v=TMZi25Pq3T8 https://www.youtube.com/watch?v=rtokNN1HZ9A

7: https://www.vice.com/en/article/amen-breakbeat-fundraiser/

8: https://www.musicbusinessworldwide.com/why-a-fight-over-61000-recordings-could-shape-the-future-of-ai-music-licensing/

9: https://www.musicbusinessworldwide.com/sony-music-files-new-lawsuit-against-ai-platform-udio-asserting-over-30000-sound-recordings-a-judge-barred-it-from-adding-to-its-original-case/

10: https://www.bbc.co.uk/news/articles/cjdrl7lr039o

11: https://www.hollywoodreporter.com/music/music-news/how-many-musicians-use-ai-1236616294/

12: https://aristake.com/ai-tools-musicians-study/

13: https://www.smithsonianmag.com/smart-news/ai-music-is-already-here-to-protect-human-artists-the-record-industry-proposes-labels-for-it-like-those-for-explicit-lyrics-180989128/

14: https://www.ethanhein.com/wp/2011/the-amen-break/