27.01.2025: UK Music’s Florian Koempel (pictured left) and Dougie Brown (pictured right) explain the details in this Q&A:

Q: The Government has launched a Consultation on copyright and artificial intelligence.  What is it about and why does it matter? 

A: The Government launched a consultation exercise just before Christmas on a variety of issues concerning copyright and artificial intelligence (AI) with a deadline of 25 February 2025 for responses. The consultation covers a number of key issues for the UK music industry including transparency, technical standards, contracts and licensing, labelling at the input side, as well as the copyright status of AI-generated works and the protection of the personality rights of individual creators against digital replicas – often known as deepfakes – at the output side.

Q: It feels that the new Government, and the previous administration have been talking about copyright and AI for some time. Why is the music industry worried? 

A: In this consultation, the Government has put forward their preference for a text and data mining exception (TDM) which will specifically benefit AI developers at the expense of human creators in the music industry and across the whole creative sector.                                                                                                                                                                                                                                                                                          They also seem to already have accepted their preferred option in the AI Opportunities Action Plan.

Q:  What do you mean by a text and data mining exception? What has mining to do with music? 

A: Not a lot. Text and data mining exceptions were introduced to permit the copying of text to access useful data; those data in themselves are not protected by copyright, and it was justified to make copies of copyright protected text to get to these data for the greater good. This is completely different from copying music in data format for the benefit of a select group of companies.

But whatever we call it, there is no need for an exception in the first place. The market is working as it is. Rightsholders have been licensing AI developers to use their repertoire for over a decade. There is no market failure which needs be corrected. In recent months, and in particular during discussions on a proposed code of practice, the music industry has been engaging with AI developers about what they actually need in addition to a licence, but to no avail. As the philosophers Jagger and Richards said: “You can’t always get what you want, but if you try sometimes, you just might find, you get what you need.”

In addition, the copyright framework is clear. Copying works to train AI models requires a licence from the relevant rightsholders unless an exception applies – which is not the case in the UK.

Q: Surely the Government will implement safeguards to protect our creative industries – which generate £125 billion annually – against misuse by global tech giants?

A: The Government has stated its intention to protect the creative industries. It would have been surprising if not, given our sector’s economic importance and the soft power it gives the UK across the globe. The Government acknowledges that the required protections don’t exist yet, and that they will not introduce an exception without effective guardrails. These guardrails need to work in practice and not be a pretext for a wide exception, window dressing. Creators should be the judge of whether these guardrails operate efficiently, protecting their creative talent.

Q: Can you explain the Government’s plans for AI in a nutshell? 

A: The Government wants to limit the rights of individual creators and rightsholders for the benefit of AI developers training their computers on the work of talented human creators without their consent and without paying them for the fruits of their labour. To make this reduction of copyright, a human right, more palatable the Government is proposing the concept of rights reservation. Essentially, creators and rightsholders would have to reserve their rights to be protected from this wide exception.

If this sounds rather peculiar, that’s because it is. It turns the copyright system upside down. If you want to keep your rights, you need to reserve your rights – otherwise an AI bot can just use your work at their pleasure. The success of our creative industries is predicated on a strong copyright framework, and any such formality such as a rights reservation is not allowed under international copyright law. Additionally, any such exception in the first place infringes all three steps of the internationally binding three-step test. The three-step test essentially limits the ability of a Government to introduce exceptions which are too wide in scope, and prevent commercial exploitation, e.g. through licensing.

Q: Isn’t the Government’s proposal similar to the opt out system in the European Union?

A: Yes, it is. So, why do we want to copy a European Union concept which has failed creators,  rightsholders and AI developers alike? The situation in the EU created substantial uncertainties for everyone involved. In general, nobody knows how to opt out (most creators don’t even know that they have to opt out). Is a general prohibition of web crawling in the Ts & Cs of one platform sufficient? Is it sufficient that creators and rightsholders publish a machine readable opt out declaration on a widely accessible website, or do they have to write a personal letter to every AI developer, existing and future? Surely, that’s absurd? Several legal cases, for example, in Germany and the Netherlands are looking into this. So are EU policy makers drafting codes of practice, but nobody knows anything for sure. Computer specialists are working on solutions but an effective technological solution protecting the rights is years away, if possible at all.

Q: Could the Government’s plan for a system of rights reservation work? 

A: A viable system of rights reservation would need to be easily accessible – not just for large rightsholders but also for individual creators. This is currently not the case, as the evidence from the European Union shows. It needs to be up-to-date and apply to all AI developers whatever the technology of reserving the rights and in a machine-readable format. The AI developers need to check the webpages they are scraping for illegal material including reserved rights. It should be the responsibility of the AI firms.

AI firms must also respect contractual reservations by platforms and personal data of individual creators. Rights reservations need to reference individual works as well as larger repertoires of rightsholders such as Collection Management Organisations (CMOs). Crucially, the system of rights reservation should only apply in regard to machine learning, and not prevent the general online accessibility of creative works. The technologies need to operate at a granular level otherwise creators and rightsholders risk being caught between a rock and a hard place, forced to decide whether to use their works commercially online and thus not declare a rights reservation.

Q: Wouldn’t that leave creators facing an impossible choice?

A: Yes, and there is no need for it in the first place.  It is far from clear why the Government has put out this consultation with its deeply flawed assumptions.

What we need is far greater transparency. AI developers should provide information on what training data has been copied; which data collection methods and sources have been used (e.g. specific data sources or general web-scraping); and a list of the specific works (songs, books etc.) used during a certain time period. This is not rocket science for tech innovators. And as an instrument of consumer protection, AI-generated music should be labelled as such to distinguish it from the work of human creators. After all, this is what human consumers want to know.

Q: What is the message the UK music industry wants the Government to hear when it comes to AI? 

A: The Government’s current proposals only offer the prospect of transparency from AI companies in exchange for a text and data mining exception to copyright. An exception would water down copyright, negate the need for licences and deal a devastating blow to the UK Music industry and the 216,000 jobs the sector supports.

The plans would seriously undermine the music industry’s efforts to grow the £7.6 billion it contributes to the economy and play its part at the heart of the Government’s industrial strategy.

The Government must urgently think again and support music and our creative industries which are such a vital part of our cultural fabric and our influence across the world.