Brussels Tinkers with the Gears of Creativity: The AI Copyright Consultation Is Here

The European Commission has officially opened the floor to what promises to be a high-stakes debate over the future of the creative economy. On May 18, 2026, the Commission launched a call for evidence to reassess the effectiveness of the 2019 Directive on Copyright in the Digital Single Market. While the review technically covers the broader health of the digital marketplace, everyone knows where the real fire is: generative AI. As developers continue to scrape the open web to fuel ever-hungrier large language models, Brussels is finally looking to see if the current "opt-out" system for text and data mining is actually doing its job or if the legal plumbing needs a complete overhaul.

It's a classic Brussels maneuver—the "targeted consultation"—but the implications are anything but narrow. The Commission is specifically hunting for feedback on how AI-driven licensing and enforcement are playing out in the real world. We've seen a wave of private partnerships and licensing deals over the last year, but for every high-profile deal between a tech giant and a publisher, there are thousands of smaller creators feeling left in the dust. This consultation, which stays open until June 25, 2026, is the Commission's way of asking whether we need a "targeted legislative initiative" to fix the power imbalance between those who build the models and those who provide the data they're built on.

The Training Data Trap

At the heart of the matter is the "Text and Data Mining" (TDM) exception. Under current EU rules, companies can generally scrape content for AI training unless a rightsholder explicitly says "no" in a machine-readable format. It's a system that favors those with the technical chops to implement complex digital handshakes. Many in the creative sector argue this "opt-out" approach is fundamentally backwards, forcing authors to play a perpetual game of whack-a-mole against an industry that’s already scaled past them. The Commission is now asking if this framework has actually fostered a "fairer copyright marketplace" or if it’s just provided a convenient loophole for Silicon Valley to digest Europe’s cultural heritage for free.

Licensing, Piracy, and the "Fair Share"

Beyond training data, the consultation is poking around in some of the more painful corners of the industry. This includes the persistent headache of online piracy—specifically for live events—and the perennial struggle for performers and music producers to get a decent cut when their work is streamed or broadcast. There’s even talk of introducing "material reciprocity" in remuneration, a move that could shake up how royalties flow across borders. While the AI Act has already set some transparency goalposts, this copyright review is where the financial rubber meets the road. It’s about more than just knowing what data was used; it’s about deciding who gets a check at the end of the month.

Stakeholders across the board, from individual artists to the biggest tech labs, have until late June to weigh in before the Commission starts drafting the next chapter of European digital law. For a continent that prides itself on being the world's "regulatory superpower," this isn't just about cleaning up the 2019 rules; it's about defining the price of creativity in an automated age.

The Granular Fight Over Metadata and Machines

Behind the Scenes: The technical tug-of-war in Brussels isn't just about whether AI can use copyrighted works, but how creators can practically enforce their refusal. The 2019 Directive introduced a machine-readable opt-out, yet years later, the industry lacks a unified standard that works across every platform and file type. While big publishers might have the engineering staff to implement complex robots.txt files or proprietary headers, the independent photographer or local journalist is often left shouting into a digital void. This consultation is the Commission’s silent admission that a right without an easy way to exercise it isn't much of a right at all.

Stakeholder camps are already entrenching themselves for a long winter of lobbying. On one side, the tech coalition argues that any move away from the current opt-out system toward a mandatory "opt-in" model would essentially kill European AI innovation before it finds its feet. They contend that the sheer logistical nightmare of clearing rights for billions of data points would leave the continent permanently trailing behind the US and China. For them, the TDM exception was a hard-fought compromise that shouldn't be dismantled just as the technology is starting to mature.

Conversely, the creative guilds see this as a once-in-a-generation opportunity to correct a massive value gap. They argue that "machine-readable opt-outs" place an unfair technical burden on the victim of the scraping rather than the scraper. There is a growing push for a "remuneration right"—a system where, even if an opt-out isn't exercised, AI labs would contribute to a collective fund that pays out to creators based on the volume of their work utilized in training sets. It’s a model inspired by old-school blank tape levies, modernized for the age of the neural network.

What many observers overlook is the historical irony at play here. When the 2019 Directive was first debated, the focus was almost entirely on "link taxes" and YouTube’s upload filters. Generative AI as we know it today was a blip on the radar of most policymakers. Now, those legacy rules are being stretched to cover a technology they weren't designed to handle. The Commission is effectively trying to perform open-heart surgery on a law that hasn't even finished its recovery from the last round of implementation.

There is also the thorny issue of "fair use" vs. "fair remuneration," a distinction that often gets lost in trans-Atlantic headlines. Unlike the United States, where the courts are currently deciding if AI training constitutes fair use, Europe has a more rigid, exception-based system. This gives the European Commission the power to be far more prescriptive. By launching this consultation, they are signaling that they aren't content to let judges decide the fate of the creative economy; they want to write the price list themselves.

Ultimately, the outcome of this call for evidence will dictate whether Europe becomes a "creative sanctuary" or a "data desert." If the rules become too restrictive, AI development might migrate elsewhere, taking the future of software engineering with it. But if the rules remain as they are, the very culture that makes Europe a global powerhouse may find itself cannibalized by the tools meant to augment it. The deadline in June is just the starting gun for a legislative marathon that will redefine the value of a human idea.

The Regulatory Paradox of the Digital Age

Reading Between the Lines: The European Commission’s consultation is built on the shaky assumption that a perfect "licensing market" is even possible for a technology that consumes data at the scale of the entire internet. Brussels seems to believe that with enough paperwork and digital handshakes, we can reconcile 18th-century concepts of individual authorship with 21st-century statistical modeling. It is a noble pursuit, but it ignores the fundamental contradiction of generative AI: the value of a model lies in its aggregate knowledge, yet the cost is being calculated at the level of the individual pixel or sentence.

There is a measured skepticism to be found in the Commission’s sudden interest in "material reciprocity." For years, European regulators have complained about Silicon Valley’s "move fast and break things" ethos, yet this consultation suggests a desire to build a walled garden that might actually stifle the few homegrown AI champions Europe has. By demanding strict remuneration for every scrap of data, the EU risks creating a landscape where only the wealthiest tech conglomerates—the very ones Brussels usually seeks to curb—can afford the entry fee. Smaller startups, lacking the legal departments to negotiate thousands of individual licenses, may find themselves locked out of their own market.

Furthermore, the fixation on "opt-out" mechanisms feels like a solution for a world that no longer exists. Even if a creator successfully signals a refusal today, their work has likely already been ingested, processed, and "remembered" by dozens of models currently in production. Retroactive enforcement is a technical nightmare that no one in the Berlaymont building seems ready to address. We are essentially asking for a way to un-bake a cake to retrieve the eggs. The implication is that we aren't just looking for a copyright fix; we are witnessing a desperate attempt to retroactively apply friction to a frictionless technology.

One also has to wonder if the Commission is inadvertently creating a "tattletale" economy. By incentivizing creators to hunt down their work within the black-box outputs of AI, we may see a surge in algorithmic litigation that benefits no one but the lawyers. The focus on transparency in the AI Act was a good first step, but transparency without a clear path to compensation is just a front-row seat to one's own obsolescence. This consultation is the sound of the gears of bureaucracy grinding against the speed of light, and the friction is starting to produce more heat than light.

Ultimately, the most likely outcome isn't a grand utopian balance between man and machine, but a messy, fragmented reality of regional blocks. If Europe tightens the screws too hard, it becomes a museum of human-made content that AI companies simply avoid. If it leaves them too loose, it betrays the very creators it claims to protect. The Commission is walking a tightrope where the rope is made of code and the safety net is made of old law books, and everyone is waiting to see which one snaps first.

Attempts to regulate AI copyright feel remarkably like trying to install a speed limiter on a lightning bolt using a very expensive piece of parchment and a quill.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn