From Dependency to Fairness: Legal Pathways for Content Creators in Platform Markets

Over the past decade, a striking paradox has emerged in digital content markets. On the one hand, we have an unprecedented explosion of creativity; independent journalists, podcasters, musicians, game developers and influencers can, in principle, reach global audiences from a laptop. On the other hand, a shrinking number of platforms increasingly determine which content is found, how it is monetised, and on what terms creators can access their audiences.

This paradox is now a central concern for policymakers and regulators. Yet, for many creators and their organisations, the legal landscape can feel opaque and fragmented. What are the real levers for change? When should we rely on competition law, and when is regulation a better route? How does the arrival of generative AI and foundation models change the picture?

This blog offers a structured way to think about platform power and maps the expanding legal toolkit that creators can use to contest and reshape it.

1. Why digital content markets concentrate

To understand which laws might help, we first need to understand the environment in which creators operate. Digital content markets are not just “competitive markets with a tech twist”. They have structural features that make concentration and dependency almost inevitable.

Several forces pull in the same direction.

Network effects: Platforms connect multiple sides of the market: users, creators, advertisers, app developers, merchants. Each additional user makes the platform more valuable to others, so scale feeds on itself. Smaller rivals find it hard to attract enough users so that they can become viable alternatives that content creators can use.

Data advantages: Large platforms accumulate vast amounts of behavioural (including interaction) data. That data is used to improve ranking, recommendation and advertising tools, which in turn attract more users and advertisers. Creators who want to be discovered have little choice but to play within these ecosystems.

Switching costs and lack of multi‑homing: Both users and creators face practical and psychological barriers to moving elsewhere: re‑building audiences, learning new tools, losing access to platform‑specific features. Even when rivals exist, going “all in” on more than one is rarely realistic.

Behavioural design and nudging: Choice architecture – default settings, recommendation carousels, notification design – quietly pushes users towards certain services and content. That design can entrench incumbents without ever explicitly banning alternatives.

Opacity and weak legacy enforcement: For years, antitrust and merger control were slow to grasp how multi‑sided digital markets worked. Many mergers were cleared, and abusive practices were tolerated or addressed only after long investigations. That allowed entrenched ecosystems to form.

For creators, the result is familiar: they find themselves heavily reliant on a handful of platforms for visibility and monetisation, with little realistic ability to walk away. It is this structural dependence that legal tools must confront.

2. Creators’ double relationship with platforms

A second conceptual step is to understand that creators do not stand in a simple “supplier–customer” relationship to platforms. They both compete with platforms and depend on them.

On the competition side, platforms increasingly move into content markets themselves. For example:

• Apple offers music streaming and radio‑like services that compete directly with music services and audio publishers.
• Amazon is not only a marketplace but also a book publisher and producer of its own content.
• Google is both a distributor of news and an advertising intermediary around news content.

At the same time, these very platforms are indispensable intermediaries:

• App developers need mobile operating systems and app stores to reach users.
• News publishers rely on search engines and social networks for traffic and on ad‑tech stacks for advertising revenue.
• Merchants and independent brands depend on large marketplaces for reach and logistics.

The platform can therefore act simultaneously as distributor, competitor, gatekeeper, data controller and quasi‑regulator through its terms and conditions. These different roles generate a recurring set of problems for creators:

Brand and audience disintermediation: Platforms deliberately keep users inside their own branded environment. They may downplay or hide the logo, name and other distinctive features of the underlying provider, so that content appears as “platform content” rather than the work of a specific creator or publisher. At the same time, platforms collect detailed behavioural and audience data but rarely share it in a way that would allow creators to understand, reach and nurture their own communities. The result is that the platform, not the creator, owns the user relationship.

Sherlocking: Platforms monitor what works on their services and then copy or absorb successful third‑party ideas, from app functionality to content formats.

Unfair terms and conditions: Commission levels, ranking rules, bundling, and opaque review processes can leave creators and business users with little meaningful choice.

Anti‑steering and data hoarding: Rules that prevent creators from directing users to cheaper or alternative offers, and refusals to share audience data, make it hard to build independent business models.

Self‑preferencing and opacity: Platforms can systematically favour their own services, content or ads in ranking, while external actors struggle to see or contest the criteria.

Copyright exploitation: From legacy issues around news snippets to the use of creative content in training AI models, platforms may rely on creator outputs in ways the law did not originally anticipate.

These problems are interconnected and cannot be addressed adequately through a single legal lens. That is why it helps to think in terms of a toolkit rather than a single silver bullet.

3. Mapping the legal toolkit: a layered approach

The modern regulatory landscape for digital platforms is messy but not incoherent. We can think of it along three axes:

1. Horizontal vs sector‑specific: Some instruments, such as general competition law, apply across all sectors. Others target particular domains, such as media, copyright or online intermediation.
2. Transparency vs beyond transparency: Some laws mainly require platforms to explain what they are doing – disclosure of ranking parameters, reasons for suspension, data use. Others go further and impose hard conduct rules: do not self‑prefer, do not impose certain clauses, must share certain data.
3. Symmetric vs asymmetric: Some rules apply to all players, including smaller intermediaries. Others are asymmetric, aimed only at firms with significant market power, such as dominant undertakings or “gatekeepers”.

For creators and their organisations, the strategic question is which combination of these instruments can best address particular harms. In the rest of this post, I sketch two broad families of tools: competition law and regulation.

4. Competition law: from price to power, and from markets to ecosystems

4.1 The promise and limits of antitrust

Competition law has been the main legal response to platform power for many years. It offers four principal levers: rules on anti‑competitive agreements, abuse of dominance, merger control and, in the EU, State aid. In practice, most high‑profile platform cases have involved abuse of dominance or merger review.

Enforcement, however, has struggled with the realities of digital markets. Traditional competition analysis was often built around price as the key parameter of consumer welfare. In markets where users “pay” with data, attention or lock‑in rather than money, that framework is incomplete. Moreover, abuse cases are inherently ex post: they can only address conduct after harm has already materialised, and investigations often last many years. Remedies are frequently modest compared to the structural advantages they target.

Recent enforcement has corrected some of these flaws, but the built‑in limitations of case‑by‑case antitrust remain. This is one reason why regulators have turned to more proactive regulation. Still, competition law continues to matter – especially through a series of “path‑breaking” platform cases, such as Google Shopping, Google Android, and the German Facebook decision. More recently, we see the role that competition law can play as a bargaining tool for copyright and AI-related content use, an issue that has become a particularly pressing concern in the creative industries. This is a new frontier I wanted to discuss in more detail.

4.2 Competition law as a bargaining tool for copyright and AI-related content use

The French antitrust proceedings against Google in relation to neighbouring rights for press publishers remain an important illustration of this dynamic. When France transposed into national law the DSM Copyright Directive, which establishes the so-called neighboring right for press publishers, Google informed publishers that indexing/display on Search Search was conditional on publishers granting free licences. This prompted the French competition authority to intervene, imposing interim measures, fines for non-compliance and binding commitments on Google to negotiate in good faith and provide the information necessary for meaningful bargaining.

That line of enforcement is now especially significant because the competition issues around content use have moved beyond snippets and neighbouring rights into the field of generative AI. In December 2025, the European Commission opened a formal antitrust investigation into whether Google breached EU competition rules by using online content to support AI-powered services on its search results pages. The Commission stated that it was concerned that Google may have used the content of web publishers to provide these generative AI services without appropriate compensation and without offering publishers an effective possibility to refuse such use.

This development is important for at least three reasons.

First, it frames the problem as one of unfair trading conditions imposed by a dominant player. The theory is not limited to classic copyright infringement, nor to a refusal-to-deal scenario. Rather, it focuses on whether a dominant search platform can appropriate creators’ content to power its own downstream AI features while depriving publishers of a meaningful choice between inclusion in search and exclusion from AI uses.

Second, the case highlights how control over ranking, visibility and user traffic can be leveraged to extract content on terms that publishers cannot realistically reject. This recalls the logic of the neighbouring-rights disputes: when a platform is an unavoidable route to audiences, an apparent “choice” may be commercially illusory. In the AI context, the issue becomes even sharper because the platform is no longer merely displaying or indexing third-party content; it is using that content to generate synthetic answers that may keep users on the platform and reduce referral traffic to the original publisher.

Third, the investigation has broader significance for creators beyond the press sector. The Commission’s investigation is not confined to traditional web publishers but also touches on the use of online content more broadly, that is, creator content hosted on YouTube. That matters because the same structural concern arises across the digital economy: a large platform may use third-party creative output both as an input into its own AI products and as a competitive substitute for the creators’ own distribution and monetisation channels.

The case sits at the intersection of several legal regimes. Copyright law asks whether protected material may be used, licensed or reserved for AI-related purposes. The AI Act is beginning to require greater transparency around training data. But competition law adds something different: it asks whether a dominant platform is using its market position to impose unfair conditions, including to deny effective opt-out mechanisms and to shift value from content creators to its own AI interfaces in a way that distorts competition in adjacent markets.

In that sense, the Commission’s Google investigation may become as important for the AI era as the French neighbouring-rights proceedings were for the earlier conflict over snippets and news previews. If the Commission pursues the case vigorously, it could establish that competition law is not merely a back-up to copyright, but an autonomous route for challenging the way dominant digital intermediaries monetise and internalise third-party content in generative environments.

4.3 Verdict on competition law

The cumulative lesson is not that antitrust is irrelevant. To the contrary, it has helped reframe how we think about market power in zero‑price, multi‑sided digital markets, and it has pushed toward more structural solutions in areas like ad‑tech.

But the basic limitations remain:

• decisions are firm‑specific, not generally applicable rules;
• the burden of proof for abuse is high; and
• remedies, even when ambitious on paper, can be undermined in practice.

For creators looking for timely and more predictable change, these constraints are significant. That is why policymakers have increasingly turned to ex ante regulation.

5. Regulation: from transparency to hard rules

5.1 Transparency as a baseline: the P2B Regulation

One of the earlier regulatory responses to platform–business user imbalances was the EU Platform‑to‑Business (P2B) Regulation. It is a horizontal, largely transparency‑focused instrument that applies to online intermediation services and search engines regardless of their market power.

For creators and other business users, P2B offers some useful hooks:

• clearer rules on platform terms and conditions;
• rules on brand attribution;
• obligations to explain suspension or termination of accounts;
• transparency around the main parameters determining ranking; and
• internal complaint‑handling mechanisms.

The P2B Regulation does not, however, fundamentally rebalance bargaining power. It may make it easier to see that a practice is unfair, but it does not necessarily prohibit it. That is where newer, more asymmetric instruments come into play.

5.2 The Digital Markets Act: moving upstream

The Digital Markets Act (DMA) represents a decisive shift. Instead of asking, case by case, whether a platform is dominant and whether a specific practice is abusive, the DMA identifies a set of gatekeepers and imposes a list of pre‑defined obligations and prohibitions on them.

Several of these are directly relevant to creators:

• Prohibitions on using data generated by business users to compete against them (“Sherlocking”).
• Requirements to share certain audience and performance data with business users and advertisers.
• Restrictions on combining personal data across services without consent.
• Portability obligations.
• Transparency duties toward publishers and advertisers.
• Prohibitions on anti‑steering practices that prevent business users from directing end users to alternative purchasing channels or offers.
• Requirements to grant fair, reasonable and non‑discriminatory access to app stores and search engines.

Crucially, once an undertaking is designated as a gatekeeper, these obligations apply across the relevant core platform services, without the need for lengthy dominance assessments in each sub‑market. Non‑compliance can trigger significant fines and, in extreme cases, structural remedies.

For creators, the DMA opens new avenues:

• Anti‑steering obligations can underpin demands to communicate directly with users about alternative pricing or distribution channels.
• Data access and portability rules can support efforts to build independent relationships with audiences and advertisers.
• Anti‑Sherlocking provisions can help challenge copying and unfair use of business user data.

The effectiveness of the DMA will depend on enforcement, which is only just beginning. But it already shifts the bargaining baseline.

5.3 Copyright reform and AI: the DSM Directive and AI Act

Alongside platform‑specific rules, copyright law is being reshaped to respond to new uses of content.

The DSM Copyright Directive introduced, among other things, a text‑and‑data‑mining (TDM) exception. The scope of the DSM Directive’s commercial TDM exception is ill‑suited to generative AI. Properly interpreted, Articles 3 and 4 are tailored to automated analytical techniques that “analyse” text and data to extract information such as patterns, trends or correlations, not to systems that internalise and recombine the expressive features of protected works to produce new outputs. Generative foundation models therefore sit uneasily within the legal and conceptual contours of TDM and should not be regarded as automatically sheltered by the exception. Against that background, framing large‑scale training of generative models on copyright‑protected content as mere text and data mining risks diluting rightholders’ exclusive powers beyond what the legislature intended and undermining the careful balance struck by the DSM Directive. We are expecting both case law and the upcoming revision of the DSM Copyright Directive to address these issues.

The AI Act adds another layer, particularly for providers of general‑purpose AI models. It requires them to publish a “sufficiently detailed summary” of the copyright‑protected works used for training, intended to help right holders understand and, where appropriate, enforce their rights. Whether these summaries will be genuinely actionable or remain highly aggregated remains to be seen. For creators and their associations, engaging with the implementation of these provisions – including templates and guidance – will be critical.

5.4 Media pluralism and user interfaces: the European Media Freedom Act

Finally, the European Media Freedom Act (EMFA) addresses yet another dimension: media pluralism and editorial independence in a platform‑mediated environment.

Among other things, the EMFA:

• imposes safeguards for editorial independence and media freedom;
• requires procedural guarantees when platforms suspend or restrict media content, including transparency and due process; and
• obliges Member States to assess certain media‑related concentrations with a view to their impact on pluralism.

It also touches on user interfaces, including rights to change default settings that control access to audiovisual media services.

These provisions may feel distant from day‑to‑day creator concerns, but they reflect a growing recognition that interface design and default choices are themselves regulatory levers. As more content is consumed through curated feeds, personalised home screens and AI‑generated overviews, the line between infrastructure and editorial decision‑making blurs.

6. What should creators and their organisations do?

The expanding legal toolkit can be overwhelming. But several strategic implications emerge.

6.1 Think in layers, not silos

No single instrument – not competition law, not the DMA, not copyright or the AI Act – will solve all problems. Creators and their organisations should think in layered strategies:

• Use competition law for structural abuses, particularly where conduct entrenches gatekeeper control or forecloses rivals.
• Rely on ex ante regulation (DMA, P2B, EMFA) to secure day‑to‑day rights around data access, anti‑steering, ranking transparency and procedural fairness.
• Deploy copyright and neighbouring rights, supported when necessary by competition enforcement, to improve bargaining positions over content use, including for AI training.

The most promising cases will often sit at the intersection of these regimes, not neatly within one of them.

6.2 Engage early with rule‑making and enforcement

Many of the most consequential choices are now being made not in abstract legislation but in secondary rules, guidance and enforcement practice:

• How regulators interpret anti‑steering and data access under the DMA.
• How TDM reservations and AI training summaries are operationalised.
• How media pluralism assessments under the EMFA interact with broader competition concerns.

Creators, publishers, collecting societies and unions should play an active role in consultations, standard‑setting and strategic litigation. Waiting until the dust settles risks locking in interpretations that privilege platform convenience over creative autonomy.

Alliances across sectors can strengthen the case that certain patterns are systemic, not idiosyncratic.

6.4 Rethink success: beyond fines

Finally, it is important to be realistic about what “success” looks like. Fines, even large ones, rarely change market structure on their own. More promising outcomes may include:

• enforceable commitments that change ranking, access or data‑sharing rules;
• structural remedies that separate conflicting functions in ad‑tech or app distribution;
• implementation of DMA obligations in ways that are genuinely usable by creators; and
• practical tools – dashboards, APIs, audit mechanisms – that make legal rights actionable.

In this sense, the most meaningful victories may be less spectacular than some past headline‑grabbing cases, but more directly relevant to everyday creative practice.

Conclusion: taming big tech by design, not by accident

Digital platforms are not neutral conduits. They are institutions whose design choices shape which voices are heard, how value is distributed and which business models are viable. Law can no longer treat them simply as one category of firm among others.

For creators, the emerging legal landscape is both challenging and full of opportunity. Competition law is learning to speak the language of ecosystems and interfaces. New ex ante instruments are setting baseline rules for gatekeepers. Copyright and AI regulation are beginning to grapple with the realities of data‑hungry models.

The task now is to make these tools work together, and to ensure that the voices of creators are present when they are interpreted and enforced. The point is not to punish “big tech”, but to redesign digital markets so that creative work, in all its forms, can flourish on fairer and more sustainable terms.

The image is AI generated