The AI and copyright law policy dilemma

Posted: 24/11/2025

The advent of generative AI is transforming the world we live in, and the law is still racing to catch up. The rapid global proliferation of AI tools and expansion of the AI industry has already exposed problems with UK copyright law.

The recent case of Getty Images v Stability AI has highlighted the challenges copyright owners face in defending their rights. Getty found itself unable to rely on primary or secondary copyright infringement claims to enforce its IP rights. The outcome of this case underscores a broader dilemma for UK policymakers: how to strike a fair balance between the interests of AI developers and copyright holders. The government has several options before it. The path it chooses will have significant implications for key sectors of the UK economy.

The policy dilemma

To meet the challenge posed by generative AI, the government must weigh two distinct sets of interests:

• Creatives and the creative industry: These stakeholders are concerned with fair remuneration, transparency around how their works are used, and maintaining control over licensing.
• AI developers: This group seeks broad freedom to mine data, reduced licensing burdens, and legal certainty to support innovation and investment.

Creators argue that their works are being used without consent or compensation, while developers maintain that their models do not store or reproduce works and that overly restrictive rules could stifle progress.

What is the UK government doing?

So far, the government has set out proposals and launched a consultation framework to explore potential reforms. The government's consultation, entitled 'Copyright and Artificial Intelligence', is part of a broader legislative initiative aimed at modernising data use and access, including how copyright interacts with AI development. The consultation opened in December 2024 and closed in February 2025, receiving over 11,000 responses.

The proposals in the consultation include:

• retaining the current legal framework;
• requiring licensing for all uses of copyrighted material in AI training;
• introducing a broad exception for data mining; and
• creating an opt-out model, where content can be mined unless rights holders actively reserve their rights, potentially supported by transparency obligations requiring developers to disclose the sources of the data they use for model training.

Under the Data (Use and Access) Act 2025, the government is obliged to report on the consultation outcomes and publish an economic impact assessment by March 2026, which is likely to be the next indication we have of the direction the government will go. The creation of an opt-out model is currently the government's preferred approach.

Whatever the outcome, it is likely that the government will need to combine legislation with increased responsibilities for the Information Commissioner's Office and the UK Intellectual Property Office to achieve a workable solution.

The global context

As the UK government moves forward, it will not be taking decisions in isolation from the rest of the world. Other territories are facing the same dilemmas, and the UK government will need to consider whether harmonisation or charting an independent course will be most beneficial. These decisions have become all the more pressing in the wake of the Getty case and are of great importance at a time when the government is struggling to get the UK economy moving.

The EU is the UK's largest trading partner and has within it major national competitors to the UK's AI ambitions in France and Germany. The EU has already pressed ahead with initiating wholesale reform in relation to AI and copyright. The EU AI Act, which came into force in 2024, was the world’s first comprehensive legal framework for artificial intelligence. It classifies AI systems by risk level and imposes strict obligations on providers of high-risk and general-purpose AI models. For general-purpose AI, the act requires developers to publish summaries of training data sources and to comply with copyright rules, including respecting opt-outs which must be expressed by copyright owners in machine-readable formats.

Alongside the AI Act, the EU has introduced a directive on AI and copyright, which reinforces the right of creators to reserve their works from being mined for AI training. The directive supports transparency and lawful data use, encouraging AI developers to avoid scraping from infringing sources and to implement safeguards against reproducing protected content.

Outside of Europe, other global competitors in the AI space are charting a different course and opting against wholesale regulatory reform. For example, the USA is yet to bring in any federal legislation regarding AI and copyright and are instead relying on guidance issued by the US Copyright Office (USCO) to clarify the position. As noted by the USCO, broad fair use exemptions in US copyright law give AI developers room to manoeuvre when it comes to using copyrighted materials to train AI tools.

Where next?

When it comes to addressing the impact of AI on copyright law, the UK will soon decide which road to go down. The health of UK creative industries and burgeoning AI sector will be impacted by the strength of copyright protections and form of opt-outs that the government chooses.

These choices will be made in a global context. Greater harmonisation with the EU could simplify compliance for developers operating across borders and support international collaboration. Diverging could allow the UK to tailor its approach to domestic priorities and carve out a distinct space as an AI industry hub. Divergence, however, risks creating greater legal uncertainty for developers and friction between developers and rights holders alike if it results in broader exemptions for AI development.

At a time when the UK government is finding it difficult to generate the kind of economic growth needed to support its spending plans, striking the right balance between the interests of creatives and AI developers will be crucial.

This article was co-written by Jake Watts, associate in the commercial, IP and IT team.

Return to news headlines