Rights holders and generative AI companies: infringement risks meet competition pressure
To date, much of the legal debate concerning generative AI has centred on issues of copyright law. Policymakers, courts, and rights holders continue to grapple with the question of whether AI developers should be permitted to use copyrighted works to train their models and, if so, on what terms.
The copyright concerns surrounding generative AI are well established. Generative AI systems are commonly dependent on the use of vast quantities of data. This data allows generative AI systems to be trained. Depending on the nature of the system and training data, this material is likely to include copyrighted works. Rights holders have questioned whether this use of their works is lawful and, if it is not, whether AI developers should be required to obtain licences before using such content for training purposes.
At the same time, AI developers have argued that access to large datasets is essential to the development of effective AI systems. They contend that imposing extensive licensing requirements could increase costs, reduce access to training material, and potentially place businesses at a competitive disadvantage.
The result is an increasingly familiar policy dilemma. Rules that favour rights holders may inhibit AI development, while rules that favour AI developers could weaken incentives for the creation and protection of original works.
Much of the current debate has therefore focused on where the appropriate balance should be struck between innovation and intellectual property protection.
The UK government has recently swerved away from immediate changes to copyright legislation. However, in June 2026, the UK’s competition regulator, the Competition and Markets Authority (CMA), imposed new requirements on Google requiring it to provide publishers with greater control over the use of their content in AI-powered search features. Among other measures, publishers will be able to opt out of having their content used in AI features such as AI overviews, whilst remaining visible in traditional search results. Google must also provide clearer attribution and links to publisher content used in AI-generated results. The CMA described the measures as a ‘world first’ designed to give publishers greater control and bargaining power over the use of their content.
This is a notable development. The debate surrounding generative AI has traditionally been framed as a conflict between AI developers seeking access to large volumes of data and rights holders looking to protect and monetise their content. However, this decision by the CMA shows that competition regulators are now actively examining the relationship between AI developers and content providers in terms of transparency and market power.
As a result, generative AI companies may find themselves facing pressure from two directions: copyright owners searching for greater control over the use of their content, and competition regulators seeking greater transparency regarding how that content is accessed and used. The latter is likely to feed the former: greater transparency may improve the ability of rights holders to identify potential infringements, assess licensing opportunities, and enforce their rights.
Why competition regulators are becoming interested
This development shows that the evolution of AI markets is now firmly in the sights of competition regulators. Access to data has become a critical input for AI development, while a relatively small number of companies possess the resources, infrastructure, and distribution channels necessary to develop and deploy advanced AI systems at scale.
Consequently, regulators have begun to examine the relationship between large technology companies and content providers more closely. In particular, attention has focused on whether publishers and other rights holders have sufficient visibility regarding the use of their content, and whether they are able to make informed decisions about how that content is used within AI systems.
The CMA’s recent intervention suggests that transparency is becoming a key area of focus for the regulator. Although such measures are not intended to resolve copyright disputes, they nevertheless have the potential to influence the way in which those disputes develop.
This reflects a broader competition concern. Where a platform is one of the main ways users find information online, publishers may have limited practical choice over whether to make their content available through it. If that content is then used to generate AI responses that reduce the need for users to view the publishers’ own material, transparency becomes relevant not only to copyright enforcement, but also to the fairness of the commercial relationship between platforms and content providers.
Transparency as common ground
One notable feature of the current debate is that copyright and competition concerns increasingly point towards the same practical solution: greater transparency.
For rights holders, transparency may provide valuable information regarding whether content has been used for AI training purposes and the extent of that use. Without such information, it could be difficult to assess whether legal rights have been infringed, if licensing opportunities exist, or whether enforcement action is appropriate.
For competition regulators, transparency may address concerns regarding information asymmetries between large technology companies and content providers. Publishers may be better placed to negotiate, license, or withhold content if they have greater visibility regarding how that content is being used.
The objectives are different, but the outcome is similar. Both copyright owners and regulators are placing increasing pressure on AI developers to provide greater visibility into the acquisition and use of content.
What does this mean for AI developers?
Greater transparency, monitoring, and content management obligations may provide meaningful benefits to rights holders. However, implementing such systems at scale is unlikely to be straightforward or inexpensive.
This raises an important question for policymakers. Measures designed to protect rights holders and improve transparency may also increase compliance burdens for AI developers. Those burdens could be more easily absorbed by large, established technology companies than by smaller competitors seeking to enter the market.
The challenge therefore extends beyond balancing the interests of AI developers and rights holders. Policymakers must also consider the potential impact of any regulatory measures on competition within AI markets.
In that sense, generative AI companies increasingly appear to be caught between competing pressures. Copyright owners are seeking greater control over the use of their content, while regulators are looking for greater transparency and accountability in relation to AI systems.
What does this mean for rights holders?
For rights holders, the trend towards greater transparency may prove significant regardless of how the substantive legal questions surrounding AI training are ultimately resolved.
One of the practical challenges facing copyright owners has been the difficulty of determining whether their works have been used for training purposes and, if so, how. The limited visibility currently available can make it difficult to assess potential claims or develop effective enforcement strategies.
Those difficulties were demonstrated in Getty Images v Stability AI. In that case, the High Court considered claims brought by Getty Images against Stability AI in relation to the Stable Diffusion model. The case highlighted the evidential and jurisdictional challenges that can arise when rights holders seek to establish where training occurred and how particular works were used.
Measures that improve transparency may therefore alter the practical landscape in which AI-related copyright disputes are fought. Greater access to information could assist rights holders in evaluating potential claims, understanding licensing opportunities, and making informed decisions about enforcement. For copyright owners, that may be valuable irrespective of the outcome of the broader policy debate regarding AI training. Whether future reforms favour rights holders or AI developers, transparency is likely to remain a key issue.
More broadly, the direction of travel is clear. Copyright law will remain central to the debate, but competition and regulatory pressures are increasingly shaping the practicalities of identifying, and potentially enforcing, copyright.