As AI expands and the “zero-click” model takes hold, publishers and creators are being squeezed. Regulation incorporating collective rights management and sovereign infrastructure is becoming urgent in order to rebalance power and protect intellectual work, warns Vincent Lorphelin, consultant and founder of Controv3rse.
A distributor in a monopoly position is plundering its suppliers in order to compete with them more effectively: that, in essence, is how digital publishers describe Google AI Overviews.
AI-generated answers now appear at the top of Google search results. They incorporate publishers’ content without authorization or compensation, enabling Google to capture a growing share of the web’s public space through “zero-click” searches: the answer is provided directly, and users no longer need—or have far less need—to visit publishers’ links.
This marks another transformation of the Internet. AI systems, and not only Google’s, are inserting themselves between the production of information and its consumption. We could speak of the Uberization of intellectual professions, except that Uber still needs drivers.
AI, by contrast, reaches users directly. The situation is closer to an oligopsony: a handful of buyers—the AI companies—facing a multitude of sellers—the publishers—much like the recurring conflicts between Lactalis and dairy farmers. The pattern is familiar: pressure on prices, the elimination of the most vulnerable players, demands for fair pricing, regulation, sanctions, and then arbitrage in favor of countries with fewer constraints.
What should be done? Wait for the courts to decide? More than a hundred lawsuits are already pitting rights holders against AI operators. Anthropic faces $1.5 billion in compensation payments to authors. The music AI company Suno has been found liable for copyright infringement, although it may appeal. It will take years before case law becomes settled.
An €890 Million Fine for Google
Legislate? Google has just been fined €890 million under the Digital Markets Act. Prepared long in advance, this regulation is only now beginning to produce effects, and even then hesitantly, as Washington threatens Brussels with retaliation.
The Draghi report points in a clear direction: regulating innovations developed elsewhere is not enough. Europe needs an industrial policy based on sovereign infrastructure. That is the ambition behind the Cloud and AI Development Act and Chips Act 2.0, designed to help Europe catch up, if possible, with the successive waves of cloud computing, computing power and semiconductors. We should not wait for the “zero-click” wave to crash over us before acting.
Which infrastructure should be prioritized? The quality of AI also depends on the quality of the content it relies on. For the moment, AI systems favor the small percentage of sources deemed most credible and up to date, which helps explain their agreements with news organizations. But premium or specialized models will also require specific, reliable resources. Getty has begun compensating photographers whose images were used to train its tools. In music, Spotify, Universal and Warner are experimenting with forms of “royalty taps.”
From this perspective, content should be thought of less like milk and more like water or oil. Strategic power lies in controlling the resource, the concessions and the meters. Those controls must not be left in the hands of the operators.
First, the resource itself: the French Senate wants to strengthen copyright protections. We must also help those who do not yet see themselves as rights holders to assert their rights. RSL Media, an initiative co-founded by actress Cate Blanchett, points to one possible approach: allowing everyone to authorize or refuse the use of their name, image or voice by AI systems.
Toward a SACEM for AI
Next come the concessions: the French and European parliaments are calling for fair remuneration. We must go further and create a collective rights-management organization, modeled on SACEM, capable of negotiating with major AI operators and distributing royalties.
Its power would not be merely legal or economic. It would also need to draw on corporate social responsibility and reputational pressure. Apple’s advertisement showing a hydraulic press crushing musical instruments, sculptures and paint pots into a tablet provoked widespread outrage.
Apple’s retreat was a reminder that the way a company treats creativity and human work directly affects its public image. It would be better to remember Tim Cook in 2018, when he denounced the “data industrial complex,” supported data protection through Europe’s GDPR, and reframed the iPhone’s communications around the right to privacy.
Finally, the meters: Microsom and Amazon are already testing marketplaces connecting publishers and AI companies. The key technology of “attribution engines,” capable of estimating how much training content contributes to an AI-generated response, is the subject of intensive research, particularly in France. What remains is to build the infrastructure needed to measure individually what is still being captured in bulk.
If anticipated today, the zero-click wave could become an opportunity to enhance the value of our cultural and cognitive capital. A coherent industrial strategy would help rebalance the relationship between creative and intellectual work, culture, the economy and AI.