Summary: The CJEU has strengthened press publishers by confirming that Member States may require platforms to negotiate, provide value-related data, and accept regulatory oversight when using press publications online.

The ruling is valuable because it tackles the real problem: not just copyright ownership, but bargaining power, transparency, enforcement, and platform pressure.

For AI debates, it does not directly decide training-data legality, but it offers a powerful template for future rules on data disclosure, fair licensing, auditability, and remuneration.

When Copyright Finally Gets Teeth: Why the Meta–AGCOM Ruling Matters for Publishers and AI

by ChatGPT-5.5

The CJEU judgment in Case C-797/23, Meta Platforms Ireland, is important because it says something very practical: a copyright right is not enough if the stronger party can simply ignore, delay, hide the relevant data, or pressure the rightsholder during negotiations. The Court has effectively said that EU Member States may build enforcement machinery around the press publishers’ right, including fair remuneration, negotiation duties, disclosure of value-related data, and regulatory supervision, as long as the publisher remains free to authorise, refuse, or grant use for free.

In easy language, this means Italy was allowed to say to platforms such as Meta: “You cannot use press publications online, refuse meaningful negotiation, hide the data needed to calculate value, or reduce visibility as pressure during the negotiation process.” The Court accepted that such obligations limit the platform’s freedom to conduct a business, but found that this limitation may be justified to create a fair copyright marketplace, let publishers recover their investments, and protect media pluralism.

The judgment does not mean every publisher automatically gets paid for every online use. The Court set clear boundaries: payment must be linked to authorisation; publishers must be able to refuse permission or grant it for free; and no payment can be required where the platform does not use the publication. The case now goes back to the Italian court to apply the CJEU’s interpretation.

Why this matters

For years, publishers have had formal rights on paper but weak leverage in practice. The problem was not only legal uncertainty; it was bargaining power. Large platforms often know far more than publishers about traffic, engagement, advertising value, substitution effects, and the economic value created by snippets, previews, indexing, summaries, recommendations, or platform-native display.

The CJEU has now recognised that imbalance. That is the real power of the judgment. The Court says providers may be required to negotiate, to avoid visibility retaliation during negotiations, and to provide the information necessary to calculate remuneration because only they may possess the data needed to assess the economic value of the use.

For publishers, this shifts the debate from “please license our content” to “the law can require a fair process for determining the value of your use of our content.” That is a much stronger position.

The most surprising statements

The first surprising point is that the Court accepts that partial EU harmonisation does not necessarily prevent Member States from strengthening implementation at national level. Platforms often argue that EU law has already occupied the field and that national additions create fragmentation. The Court’s answer is more nuanced: Member States retain discretion to make EU rights effective.

The second surprising point is the acceptance of mandatory negotiation duties. Copyright is usually framed as a private right: the rightsholder either licenses or sues. Here, the Court accepted that the State may shape the negotiation environment itself.

The third is the “no visibility reduction” logic. The Court recognised that a platform could use ranking, visibility, or discoverability as leverage against publishers. That is extremely important because modern content markets are not only about copying; they are about access, ranking, indexing, recommendation, and platform dependency.

The fourth is the Court’s recognition that data asymmetry matters. The judgment says publishers are in a weaker negotiating position because the providers hold the information needed to assess economic value. That is not just copyright reasoning; it is platform-governance reasoning.

The fifth is the link between publisher remuneration and democracy/media pluralism. The Court does not treat publisher compensation as a narrow commercial issue. It connects the economic sustainability of press publishers to the public interest in media pluralism.

The most controversial statements

The most controversial part is that the Court allows obligations that interfere with the platform’s freedom to conduct a business. Meta’s argument was, in essence, that the Italian framework went too far by giving AGCOM powers to shape remuneration and negotiation conditions. The Court did not dismiss business freedom, but it said that limitation may be proportionate when balanced against intellectual property and media pluralism.

A second controversial element is the role of the regulator. AGCOM is not merely a passive observer. The Italian framework gives it powers to set criteria, intervene in disagreement, require information, and impose penalties. The Court accepted that such public authority involvement may be permissible where it ensures effective implementation of publishers’ rights.

A third controversial point is that this ruling may encourage more national “gold-plating” of EU copyright rules. Platforms will likely argue that this creates legal fragmentation. Publishers will argue the opposite: that without national enforcement mechanisms, EU rights remain decorative.

The most valuable statements

The most valuable statement for publishers is that fair remuneration can be compatible with EU law when it is consideration for authorising online use.

The second most valuable statement is that platforms may be required to provide the data necessary to calculate remuneration. That is potentially more important than the remuneration right itself. Without data, valuation is guesswork.

The third is that platforms may be required to negotiate without reducing the visibility of publisher content during negotiations. This addresses the “take the deal or disappear from traffic flows” problem.

The fourth is that the Court accepts the idea of effective implementation. A legal right must be capable of being enforced in the real world, not only recognised in theory.

The fifth is that the Court frames publisher investment, copyright fairness, and media pluralism as connected public-interest objectives. That gives publishers a stronger policy narrative than “we want money from Big Tech.”

How publishers can use this going forward

Publishers should use this judgment as a blueprint for a more muscular licensing and policy strategy.

First, publishers and trade associations can argue for national laws that do not merely restate the press publishers’ right but make it operational. That means rules on good-faith negotiation, disclosure of usage and revenue data, independent valuation, anti-retaliation duties, regulatory oversight, and penalties for non-compliance.

Second, publishers can incorporate the judgment’s logic into platform negotiations. The commercial ask should not only be “pay us.” It should be: “show us the usage, show us the value, preserve visibility during negotiation, disclose the methodology, and agree to an auditable remuneration framework.”

Third, publishers should build better evidence systems. The judgment makes data central. Publishers need to track where their content appears, how it is displayed, whether snippets or summaries substitute for visits, how rankings change during negotiations, and how platform use affects subscriptions, advertising, licensing, and brand value.

Fourth, publishers should coordinate carefully through associations. This ruling gives industry bodies a stronger argument for collective policy advocacy. However, publishers still need to remain alert to competition-law limits when coordinating commercial terms.

Fifth, publishers should translate this into AI-era contractual language. Future contracts with AI platforms should include usage reporting, audit rights, training/inference distinctions, data-retention limits, attribution obligations, output monitoring, anti-substitution controls, and clear rules on whether content can be used for training, RAG, embeddings, summarisation, indexing, or agentic retrieval.

The important caveat for scholarly publishers

For Wiley and scholarly publishers, the judgment is highly relevant but not automatically directly applicable to all scholarly content. The DSM Directive expressly says that periodicals published for scientific or academic purposes, such as scientific journals, are not “press publications” for the purposes of the press publishers’ right.

That means scholarly publishers should not overstate the case. This is not a direct CJEU ruling that scientific journals fall under Article 15 press-publisher remuneration. But it is still strategically useful because the Court’s reasoning can support broader arguments about bargaining asymmetry, platform-held data, the need for enforceable licensing frameworks, and the legitimacy of public-interest copyright enforcement.

For scholarly publishing, the most useful lesson is not “Article 15 now solves AI licensing.” It is this: rights without transparency, data access, auditability, and enforcement architecture are structurally weak.

Does this matter for future AI debates?

Yes, but as an analogy and policy template rather than a direct AI ruling.

The judgment does not decide whether generative AI training on copyrighted works is lawful or unlawful. It does not reinterpret the EU text-and-data-mining rules. Under Article 4 of the DSM Directive, text and data mining of lawfully accessible works is subject to an exception unless the rightsholder has expressly reserved rights in an appropriate manner, such as machine-readable means for online content.

But the ruling is valuable for AI debates because it attacks one of the central problems in AI licensing: information asymmetry. AI companies often know what was collected, copied, retained, embedded, trained on, queried, or reproduced. Publishers often do not. The CJEU’s logic supports the idea that where one side controls the data needed to value use, law can require disclosure.

That has obvious relevance to future AI policy. Legislators could use similar reasoning to require AI developers to provide meaningful information about copyrighted-content use, training-data provenance, opt-out compliance, retrieval logs, output substitution, and revenue/value attribution. The same structure could support regulated negotiation frameworks between AI companies and publishers, especially where platforms benefit from content while resisting transparent valuation.

The judgment also helps reframe AI copyright debates away from a simplistic “innovation versus copyright” story. The Court’s framing is closer to: innovation is legitimate, but so is the need for a fair marketplace, investment recovery, media pluralism, and effective enforcement. That is exactly the kind of balancing language publishers need in the AI era.

Bottom line

This is a strong judgment for press publishers and a useful strategic signal for the wider content industries. It says that Europe does not have to tolerate a world where platforms benefit from protected content while hiding the economic value of that use and using market power to weaken negotiation. The decision does not solve AI copyright. It does not automatically cover scholarly journals. But it gives publishers a powerful legal and policy architecture: fair remuneration, mandatory negotiation, disclosure of value-relevant data, anti-retaliation protections, and public-authority enforcement.

For future AI debates, its greatest value is conceptual. It confirms that copyright policy can move beyond abstract rights and into enforceable market design. That is exactly where the AI licensing debate now needs to go.