The rapid rise of generative AI is forcing governments and industries to grapple with complex questions around copyright, intellectual property, and the future of human creativity. A new report from the UK’s All-Party Parliamentary Group on Music dives into these thorny issues, calling for the government to craft a “pro-creative industries AI Bill” that would introduce new rights and obligations for AI developers using copyrighted music.
Everything you need to know:
✓ Report recommends an AI bill to protect copyright and creators in the UK music industry
✓ Tensions exist between labels wanting to license music for AI training and artists wanting a say
✓ Principles-based regulation could help serve the public good while encouraging creativity and innovation
Recommendations to protect artists from unauthorized AI use
The report, researched and funded by UK music industry body UK Music, builds on their July 2023 wishlist with eight specific policy recommendations. Key among them is the proposal for an AI Bill that would require things like transparent labelling of AI-generated content, record-keeping of datasets used to train AI models, and express permission from rightsholders before using copyrighted material.
It also recommends clarifying that AI-generated works do not qualify for copyright protection, and establishing a new “personality right” to cover the voice, image, name and likeness of artists. The report takes an international view as well, calling for Large Language Models (LLMs) to comply with UK copyright law regardless of where they were developed.
An accompanying survey of UK adults found strong public support for measures to prevent unauthorized use of artists’ work in training AI (80%), to label AI-generated music (83%), and concern that AI could replace human creativity (66%).
The debate at the heart of AI music – labels vs. artists
The tension at the heart of the AI music debate is apparent. Labels and publishers want to be able to strike licensing deals for AI training on their catalogues – deals that could be lucrative. But artists and creators want to have a say in those deals, arguing their music shouldn’t be licensed for AI without their permission. The report hints at supporting the creators on this point.
According to Kevin Brennan MP, chair of the parliamentary group behind the report, the government’s initial “tech-bro” instinct towards a completely deregulatory approach to AI was misguided. He argues it risks undermining the UK’s highly successful music and creative industries, which are major contributors to the economy.
If AI developers have already ingested copyrighted music catalogues without permission to train their models, applying new regulations retroactively could prove impossible. The “horse has bolted,” as Brennan puts it. In that case, a levy on AI music companies may be necessary, with the proceeds distributed to support up-and-coming musicians and creators.
Ultimately, politicians will need to weigh the incredibly complex arguments and chart a path forward, not based on absolutes of right and wrong, but on serving the greater public good.
“Things like encouraging human creativity, rewarding those who produce interesting and original work, and making sure that they’re able to be rewarded for that.”
Kevin Brennan MP, chair of the parliamentary group
The right balance will be tricky to strike. Overly restrictive regulation could stifle promising new innovations in AI-assisted music creation and discovery. But a complete free-for-all could devalue human artistry and undermine the livelihoods of professional musicians. Flexible, principles-based rules, clearly reflecting the intent of legislators, may be the wisest approach.
One thing is certain – the transformative impact of artificial intelligence on the music business has only just begun. Tough questions around intellectual property, fair compensation for creators, and the boundaries between human and machine creativity will only grow more pronounced in the years ahead. Policymakers who start grappling with them now will be ahead of the curve.
