ABBA’s Björn Ulvaeus has warned EU leaders that they’re allowing the EU AI Act’s copyright protections to be “watered down to a level where they will be barely effective”.
Speaking at an event at the EU Parliament on Tuesday, Ulvaeus said that “we must never be seduced by the false idea that, in the headlong rush to the new AI world, creators’ interests must be cast aside. That approach won’t work - not for the creative sector, not for the economy, or for culture, or even for the tech sector whose vast AI revenues, let’s not forget, derive from copyrighted creative works made by humans”.
Meanwhile, yesterday Martina McBride told US senators that AI deepfakes of her voice and likeness “goes against everything that I've built”. Speaking at a Senate Judiciary Subcommittee hearing on the NO FAKES Act, McBride warned that AI was being abused “by stealing people’s voices and likenesses to scare and defraud families, manipulate the images of young girls in ways that are shocking to say the least, impersonate government officials, or make phony recordings posing as artists like me”.
Those appearances, on each side of the Atlantic, come as creators increasingly coordinate their pushback against AI companies across jurisdictions.
According to CISAC, Europe’s copyright industries contribute nearly 7% of EU GDP and support over 17 million jobs, something that is jeopardised by tech companies’ attempts to water down legislation safeguarding the rights of creators.
In Brussels, Ulvaeus led creators including Marika, Alice Wonder and Patrick Sigwalt in demanding that the EU AI Act’s copyright protections aren’t watered down during implementation. The creators’ coalition is specifically concerned that technical documents that will inform the implementation of the act - specifically the Code Of Practice and Transparency Template - will gut its transparency and copyright compliance requirements.
In Washington, the NO FAKES Act has backing from nearly 400 artists, including McBride, LeAnn Rimes, Bette Midler and Scarlett Johansson. Tellingly, it’s also supported by YouTube, whose parent Google is investing heavily to compete in the generative AI product race.
The proposed legislation would create federal “digital replication rights” which would make voice and likeness federally protected property that can be licensed but not sold - at least during the lifetime of the individual involved. Just two days before the latest Congressional session discussing the act, US President Donald Trump signed the new Take It Down law which requires social media platforms to remove non-consensual images and video - including so-called AI deepfakes - within 48 hours of being notified about the content.
The NO FAKES Act is “the perfect next step to build on” the Take It Down Act, said RIAA boss Mitch Glazier, adding that it “provides a remedy to victims of invasive harms that go beyond the intimate images addressed by that legislation”.
This will help in “protecting artists like Martina from non-consensual deepfakes and voice clones that breach the trust she has built with millions of fans”. More crucially, Glazier added, it “empowers individuals to have unlawful deepfakes removed as soon as a platform is able without requiring anyone to hire lawyers or go to court”.
Glazier went on to argue that “artists’ voices and likenesses are fundamental to their work, their credibility and expression, and their careers”, warning that “unauthorised exploitation of them using deepfakes can cause devastating harm”.
YouTube’s Head Of Music Policy, Susana Carlos told senators that “by supplanting the need for a patchwork of inconsistent legal frameworks, the NO FAKES Act would streamline global operations for platforms like ours and empower musicians and rightsholders to better manage their IP”.
In her written testimony to the committee she added that the legislation “not only appropriately balances innovation, creative expression and individuals’ rights, but also offers a broadly workable, tech-neutral, and comprehensive legal solution”.
Back on this side of the Atlantic, the current focus is systematic transparency and copyright compliance as set out in the AI Act. However, the creators’ coalition led by Ulvaeus is concerned that what initially looked like comprehensive legislation will be rendered “barely effective” through bureaucratic sleight of hand.
The Code Of Practice gives AI companies significant input into their own regulation, while the Transparency Template risks becoming a box-ticking exercise that will provide little in meaningful protection to rightsholders and creators.
“If Europe backslides now, that would create a dangerous precedent”, warned Ulvaeus, signalling that “Europe does not value its creators or its creative sector”.
What is interesting about this dual-pronged approach from either side of the Atlantic is the timing: AI companies aren’t scrappy startups any more - they’re companies worth hundreds of billions of dollars, projecting unimaginably huge revenues.
Just yesterday, OpenAI announced a $6.5 billion all stock deal to acquire a company set up by former Apple design supremo Jony Ive, with OpenAI CEO Sam Altman claiming that the deal could generate a further $1 trillion in value for his company.
With those kinds of numbers in play, and the lobbying power that can buy, many people have said that creators’ interests will fall by the wayside as enormously wealthy tech companies buy the legislative change they need to continue to mint money.
However, the regulatory mood is shifting, and pressure is growing across multiple jurisdictions.
Creators have long known that digital-age power dynamics mean that individual artists complaining get ignored, while coordinated international pressure from aligned coalitions can get meetings with senators and commissioners.
And, seemingly, some technology companies recognise that the current free-for-all may not be sustainable, and may not be good for their own broader interests. Hence YouTube’s support for federal regulation of deepfakes, without which creator-driven platforms like YouTube may see a poisoning of the well as they are flooded with deepfakes and other AI generated content, damaging trust in audiences, and driving away advertisers.
As Labour MP Anneliese Midgley recently told Parliament - which is itself involved in a legislative pingpong battle between the interests of creators and tech companies - “How can we justify taking money away from British msuciaisn and handing it to tech firms for free?”
Tech companies have long promoted the fiction that AI development and creators’ rights are naturally aligned, and that by promoting the development of AI technology - ideally unshackled by the irritating constraints of anti-tech ideas like copyright - creators will somehow benefit through increased innovation.
Creators aren’t buying it. As Ulvaeus argues, “the vision has to be win-win for creators and the tech industry. That can only happen with legislation that truly and effectively safeguards creators”. Real win-win arrangements require actual negotiation, and fair compensation - not unilateral appropriation driven by lobbyists, dressed up as progress.
Last time around and in its previous iteration, the NO FAKES Act failed to reach a vote, despite bipartisan support. In Europe, creators face a different but equally challenging battle over how AI law is actually implemented, and whether the EU’s ethical AI framework becomes meaningful or an expensive sop to tech interests.
Whether creators can maintain coordinated pressure for long enough to secure wins remains to be seen, as is whether the broad coalitions of diverse creative industry participants can sustain a single focus without being divided and conquered by tech companies.
That aside, it feels as though the legislative tide is turning just enough that the days when AI companies could simply help themselves to creators’ work while dismissing concerns about fairness may - just perhaps - be ending.