Copyright and AI in Britain: the balancing act
It’s no secret that copyrighted creative works (including newspaper articles, novels, music, and images) are used to train generative AI models. The issues are complex, but the battle lines are clearly drawn. Creatives are lobbying governments to protect their rights, in what many see as an existential threat to the future of creativity itself. The widely publicized Declaration on AI Training, with more than 30,000 signatories, including high-profile writers, actors and academics, has brought growing public attention to creators’ perspectives on this topic.
On the other hand, AI companies are pushing for maximum freedoms so that algorithms can train on existing material to “turbocharge” innovation. Microsoft CEO Satya Nadella compared AI training to learning a subject from a textbook, arguing that companies should be given free rights to data to train their models. Like many other countries, the British government is in the spotlight as it tries to see how it can reconcile the competing interests of groups seeking to shape the legislation governing this rapidly developing region.
Partner at Powell Gilbert.
What is the current British position?
So far, Britain has taken a light-hearted approach to intellectual property issues surrounding artificial intelligence. For example, the Copyright, Designs and Patents Act 1988 (CDPA), established over thirty years ago, is still the main source of legislation in this area. The CDPA gives rights to copyright owners to prevent original creative works from being copied, distributed, or performed without permission. Although the CDPA has been amended over the years, it has not yet been updated to account for the AI era. This means that unauthorized copying of protected works as is is not permitted for the purpose of training AI models for commercial purposes. This contrasts with the position of the EU, where copying for commercial purposes is permitted unless the rights holder has opted out, and the US, where AI developers can seek to rely on the ‘fair use’ exemption.
In practice, enforcing this restriction in an AI context is challenging. For starters, it’s difficult to know if a particular work is being used without access to the training dataset used for each system. Even if it is established that the copyrighted works were used in the context of training an AI model, a rights holder must still prove that the copying of that work took place in the jurisdiction. To have any chance of success in such proceedings, it is essential that legal practitioners fully understand the technology underlying the allegedly infringing AI model. Although training data is necessarily copied initially (e.g. to RAM), in most cases, once entered, the AI model does not store a copy of the raw data. Instead, the AI’s neural network evolves in response to the training data. Without access to records, it is challenging to determine the identity of the dataset (and any protected works) used in training, although in some cases tell-tale features in the output can provide clues.
Both issues of copying and jurisdiction are at issue in the case of Getty v Stability AI, in which Getty has claimed that Stability AI has infringed its IP rights, both through the alleged use of its images as training data, and through the generated image output that the Getty watermark. The trial will take place in June 2025 and it will be interesting to see how these issues are addressed by the UK court.
Similar cases are pending across the pond, including the parallel US proceedings in the Getty case. The New York Times has filed a claim against OpenAI in the New York district courts, including a demand to destroy AI models that have used its content as training data. The outcome of these cases could have a drastic impact on the relationship between AI companies and news outlets regarding copyright.
How might British policy change?
The UK government is expected to tackle some of the controversial topics surrounding AI in the soon-to-be published Artificial Intelligence Opportunities Action Plan. This will likely propose changes to the CDPA to address the use of copyrighted works to train AI models. British Prime Minister Sir Keir Starmer said in a recent statement that the action plan will include rights for publishers to retain control over, and be paid for, content used for training. These changes have been long awaited. The issue of content creators’ rights was debated in Parliament in 2021 following a private bill initiated by Labor MP Kevin Brennan. This bill proposed rights to compensation for creators and a transparency obligation that would give authors the right to be informed about how their works are used.
The Brennan Bill is inspired by the EU Directive on Copyright in the Digital Single Market adopted in April 2019 and, in particular, by the provision on copyright contracts specifying that authors are entitled to appropriate compensation when they acquire exclusive rights for the exploitation of their works. While no progress was made on the Brennan Bill, the government committed to looking into how to secure revenue for creators and considered the issue as part of an AI and IP consultation.
The previous government’s response to the consultation, published in June 2022, stated that a new exception to copyright and databases would be introduced to allow text and data mining for any purpose, including for training AI models. Creators would not be given the right to opt-out, but would enjoy certain safeguards that would entitle them to compensation, for example through subscriptions that allow legal access to their works through a platform of their choice. The Prime Minister’s comments indicate that this remains the direction for legislation in Britain.
Given the uncertainty surrounding content holders’ rights at this stage, some major rights holders are anticipating changes in the law by choosing to enter into licensing agreements with AI companies. For example, the Financial Times in Britain, Axel Springer in Germany and Conde Nast in the US have signed commercial deals with OpenAI, each reportedly worth tens of millions of pounds per year. These deals give OpenAI access to the publications’ content, paying a flat fee for historical content plus an ongoing annual fee for new content. This trend has emerged over the past year and it is expected that other media outlets will follow suit.
What are the implications of potential reforms for both copyright owners and AI innovators?
Any proposed changes to this area of law will be of great importance to the UK IP landscape. Until now, copyright law has evolved slowly in response to changing technology. Lawmakers have preferred to make incremental changes to existing law, although many legal practitioners have called for more holistic changes. The reforms needed to address the challenges posed by AI could bring about such changes and it will be interesting to see how the upcoming AI Opportunity Action Plan addresses this issue.
In addition to copyright reforms, the EU recently passed the EU AI Act, which, among other things, regulates transparency and reporting requirements regarding AI models. This legislation is the first of its kind in the world. Post-Brexit, stakeholders will be interested to see whether or not Britain will deviate from the EU position and set a different direction for UK AI and copyright law. Be that as it may, the action plan will bring some very welcome clarity to this rapidly developing area.
We recommended the best plagiarism checker.
This article was produced as part of TechRadarPro’s Expert Insights channel, where we profile the best and brightest minds in today’s technology industry. The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing, you can read more here:
