Paul McCartney has entered one of the most important arguments in the AI copyright debate: who gets control when creative work is used to train artificial intelligence systems. The UK government is considering changes that would allow AI companies to train models on online content from creators unless those creators specifically opt out.
For McCartney, the issue is not whether AI can be useful. It is whether the people who make songs, lyrics and other creative works can keep meaningful rights over what they create when AI companies want to use that material as training data.
What the UK proposal would change
The proposal described in the source would shift the default position for online creative content. Instead of AI companies needing clear permission first, creators would have to remove themselves from the system by opting out.
That distinction matters because it changes who carries the work. Under an opt-out system, artists would need to understand that their material may be used, identify the relevant AI companies, and object to each one. Critics argue that this process is easier for data collectors than for individual creators.
The UK government has said the approach would give creators "real control" and transparency. But the music industry response described in the source is skeptical. The central question is whether control is real if artists must constantly monitor and defend their own catalogs.
McCartney’s warning centers on artists’ rights
In a BBC interview, McCartney urged the government to reconsider the proposal and strengthen protections for artists. He warned that the plan could create a "Wild West" environment in which creative works lose copyright protection.
His concern is especially focused on people at the beginning of their careers. McCartney said: "You get young guys, girls, coming up, and they write a beautiful song, and they don't own it."
That line captures the practical fear behind the copyright dispute. A song can begin as one person’s work, but if the rules around AI training are loose or hard to enforce, the value created from that work may move away from the person who made it.
McCartney also framed the issue in economic terms. "The truth is, the money's going somewhere. Somebody's getting paid, so why shouldn't it be the guy who sat down and wrote Yesterday?" he said.
The argument is not that AI has no place in music. McCartney recently worked with AI on the final Beatles track Now and Then. His position is that useful technology should not come at the expense of creative ownership.
As he put it: "I think AI is great, and it can do lots of great things. But it shouldn't rip creative people off. There's no sense in that."
Why the opt-out model worries the music industry
The music industry’s objection is partly about enforcement. If every artist has to opt out company by company, then the system depends on creators having the time, knowledge and leverage to do that effectively.
Tom Kiehl of UK Music said there is "no evidence that creatives can effectively 'opt out' of their work from being trained by AI systems and so this apparent concession does not provide any reassurance to those that work in music."
That criticism points to a gap between a legal mechanism and a workable reality. An opt-out right may exist on paper, but it may not protect creators if they cannot discover where their work is going or stop its use across different systems.
The source also notes that critics see the burden as unfair. Artists would have to track and object to each AI company individually. In practice, that can favor organizations collecting data over the creators whose work gives that data value.
Possible paths and growing legal pressure
The source points to YouTube’s recent approach as one possible model. Its system lets creators choose which AI companies can use their content. Over time, that kind of structure could lead to systematic payment for training data.
But the source also makes clear that such a system would not be simple. To work at scale, it would require coordination across platforms and countries. Without that coordination, creators could still face fragmented rules and uneven protection.
The courts are already becoming part of the wider conflict. Major record labels in the U.S. are taking legal action against AI music generators. In Germany, Gema has sued Suno.ai and ChatGPT over song lyric usage.
Those cases show that the AI copyright question is not limited to one country or one platform. The same basic tension appears wherever AI systems are trained on creative material: companies want access to data, while artists and rights holders want control, credit and compensation.
The core question is who controls creative work
McCartney’s intervention gives the debate a clear public focus. The question is not simply whether AI should be allowed in music. It is whether creators should have to chase companies to protect their work, or whether companies should have to secure permission before using it.
The answer will shape how musicians, songwriters and other creators interact with AI systems. If the default favors broad training access, creators may face a constant burden to opt out. If the default favors permission and payment, AI companies may need more structured agreements before using creative work.
For now, McCartney’s message is direct: AI can be valuable, but it should not undermine the people whose work makes creative technology possible.