Creative Gridlock: Are Rightsholders Still Fighting the Last War?
In Brussels, a recent closed-door roundtable offered a glimpse into the growing rift between AI developers and Europe’s creative industries. Hosted under Chatham House rules, the meeting centred around MEP Axel Voss’ upcoming own-initiative report for the European Parliament’s Committee on Legal Affairs (JURI) on the opportunities and challenges of copyright and generative artificial intelligence. It quickly became clear that some rightsholders are less interested in building bridges than in guarding moats.
The conversation was dominated by large players—representatives of major content industries and big tech—leaving civil society and research communities underrepresented. It is unclear what the cause of this underrepresentation was, but this skew inevitably shaped the tone, with more focus on protecting market share than on building an inclusive, innovation-friendly regulatory environment.
The session laid bare a pattern that’s becoming all too familiar: industry incumbents treating innovation as an existential threat rather than a collaborative opportunity. While AI raises legitimate questions about authorship, transparency, and licensing, many of the arguments on display were not just outdated—they risk freezing Europe’s creative sector in amber.
AI as Scapegoat: Overstating the Threat to Culture?
One of the most common refrains was that creative “works” are being reduced to mere “data,” a framing that some participants found offensive. But the real issue may be the refusal to accept that we now live in a mixed digital environment where content and data constantly overlap.
Calls to preserve “exclusive rights” at all costs often ignored the fact that many sectors already license their content and profit from digital models. Even in sectors like music and film, where structured licensing systems exist, participants doubled down on fears rather than exploring how to modernise frameworks for the AI age.
The notion that AI developers are looting culture was voiced repeatedly, yet few concrete examples were provided, and even fewer suggestions emerged for how to build viable, scalable licensing tools. Traditionally, when the debate veers toward invoking theft and cultural erosion, it becomes more about gatekeeping than governance.
The Transparency Trap: Real Problems or Rhetorical Weapons?
Rightsholders also focused heavily on transparency, especially regarding how datasets are compiled and used. There’s no doubt that a level of clarity around training data is important. But the proposed fixes often seemed disconnected from practical realities.
Demands for full URL-level disclosure, for instance, sound good in theory, but ignore the scale and complexity of training modern AI models. Pushing back against trade secret protections without offering a viable alternative may also undercut Europe’s own innovation goals. At some point, the conversation needs to shift from “what we demand” to “what can actually work.”
Some stakeholders also blamed AI companies for failing to respect opt-out signals like Robots.txt. Yet these same critics often conceded that such mechanisms are outdated and ineffective. Rather than investing in smarter tools or interoperable standards, the fallback posture seemed to be: regulate now, innovate later.
Who Gets to Opt Out? Collective Licensing’s Fractured Front
One of the more revealing moments came when participants discussed who exactly should be allowed to execute an opt-out. The question exposed deep divisions—not just between sectors, but within them—on the role of collective licensing schemes.
While some see collective management organisations (CMOs) as a natural vehicle for simplifying negotiations, others were clearly sceptical. Performers, in particular, voiced concern that collective schemes might bypass individual consent, especially when biometric data is involved. The underlying tension is clear: not everyone believes that their rights or identities can or should be pooled.
The broader implication is that building default licensing pathways may not be as straightforward as some policymakers hope. If trust in collective schemes is uneven, any system built around them may inherit that fragility.
Article 4 and the Limits of Law
Article 4 of the Directive on Copyright in the Digital Single Market, allowing text and data mining for commercial uses, came under fire from virtually all sides of the rightsholder camp. But their interpretation of its purpose and function often leaned heavily on politics and sentiment, not legal substance.
It was also hinted that Article 3 of the Directive, pertaining to text and data mining for scientific research, was currently open for abuse. Yet the argument that this exception should only apply to narrow scientific research, and not to any large-scale AI training, fundamentally misses the point. Europe cannot afford to legislate its research into a corner just because some sectors failed to keep up with technological change.
Ironically, some participants acknowledged that licensing systems are in place and functioning, especially in music, yet still argued that these don’t go far enough. If the infrastructure exists, why not use it? Blocking data access outright doesn’t protect creativity. It protects the status quo.
Research on the Sidelines: A Warning from the Margins
Amid the industry-heavy discourse, there were a few voices urging caution against regulatory overreach, most notably from the research community. One intervention stressed that any new rules must avoid collateral damage to the research sector, particularly smaller institutions and projects.
As the lone voice speaking for research reminded the room, clarity is needed on where copyright applies and where it may not be the best tool at all. Not all uses of data are exploitative. Sometimes, they’re essential for public-interest work, education, or scientific discovery. A regulatory approach that ignores this nuance risks shutting out precisely the communities Europe most needs to empower.
The contrast was stark: while major industry players focused on commercial certainty, others simply wanted room to operate without being caught in the legal crossfire.
Political Action or Policy Paralysis?
The loudest call was for “political solutions” instead of court-based ones, despite the fact that courts offer the kind of case-by-case nuance many sectors claim to need. But political intervention is only effective when it’s grounded in problem-solving, not protectionism.
MEP Axel Voss tried to steer the discussion toward solutions, urging stakeholders to come forward with pragmatic proposals. Yet many seemed more interested in re-litigating the last ten years of copyright reform than in confronting the realities of the next ten.
Performers, publishers, and filmmakers warned of cultural loss, unfair competition, and “big tech” arrogance and lack of respect. But much of this sounded like a familiar script, repeated from past debates on digital platforms, file sharing, and even social media.
A Fork in the Road for EU Policy
With timelines still fluid and more studies to come, the EU has a narrow window to get this right. The worst-case scenario is not a flawed compromise: it’s paralysis. Rightsholders are right to want fair treatment (don’t we all), but fairness is not synonymous with control.
If Europe wants to lead on responsible AI, it has to embrace flexible, forward-looking frameworks. This includes improving transparency and refining opt-out tools, but also recognising that the creative economy doesn’t benefit from fear-mongering and stalled licensing talks.
In the end, cultural value isn’t undermined by technology. It’s eroded by institutions that cling to outdated systems, reject change, and define every evolution as a threat. Europe must choose: does it want to protect models from the past, or shape the future?
Next Steps
On 12 May 2025, the European Union Intellectual Property Office (EUIPO) will present a study on generative AI and its implications for copyright. Additionally, before the summer break, the European Parliament’s Policy Department is expected to present its own study, and MEP Axel Voss hopes to be able to provide an initial preview of his draft report.
Written by Caroline De Cock, LL.M. , Head of Research.
May 9, 2025
