UK law and generative AI: food for thought

If you had told me 5 years ago that I could have a conversation about the meaning of life with an algorithm and then get it to synthesise an image of a sumo wrestler riding a crocodile, I’d have thought you were either a time traveller from the distant future, or a bit odd. 

How wrong I was. Today, AI permeates almost every level of society, and whilst it can certainly do these things, it may also be used in recruitment, weapons systems and cyberattacks. 

In this article, I will briefly explore the stance of various jurisdictions on AI, describe the recent copyright infringement claim brought by Getty Images, and discuss the importance of AI legislation for British citizens.  

Around the world, various nations have demonstrated different approaches when implementing regulations and legislation around AI. Government strategies differ around the world with some like Denmark, concerned with the social implications of deepfakes, and the misuse of people’s likeness in financial crime or fraud. In Japan, conventions concerning AI maintain a human-centred approach,(1) with this small state taking an active role in keeping dignity and respect as principal values. Singapore, has taken a less restrictive approach (2) has, well aware of the lucrative investment potential from rising tech-giants. Modern super-powers, China and the U.S., vie for AI supremacy and search for new ways to integrate autonomous systems into their military capacities. So which camp does the UK fall into? It’s little surprise that the answer is complex and confused. Contrary from what might be expected from a small state, the UK ranks as the world’s  3^rd largest AI market (3) valued in 2024 at £72.3 bn. The UK government maintains an optimistic and eager approach (4) to implementing AI into areas of public service, aiming to streamline processes and reduce costs, and has not yet passed a unifying statute about genAI and relies instead on preexisting legislation.

Inevitably, AI brings with it a host of issues. Cyberspace is a notoriously difficult area to create robust legislation for, partially because of its rapidly evolving nature. This causes further challenges when enforcing the law. An interesting case which explores this is the recent Getty Images (US) Inc & ors v Stability AI Limited [2025] EWHC 2863 (Ch).  Here, the claimants (Getty Images) alleged that the defendant (Stability AI Ltd) was infringing on copyright by training their model on images acquired from Getty without their consent. This was done outside the UK in a permissive jurisdiction, before Stability was then released in the UK. The case did find infringment of copyright, but dismissed the rest of the claim (5). and importantly made no material change to the application of English law when involving AI. However, this judgement does appear to permit the training of AI in lenient jurisdictions and then transferring them to the UK. Most importantly, the judge maintained that the findings were “both historic and extremely limited in scope” (6). This suggests that whilst the court recognises the importance of the case, it has confined its decision to the case’s facts, rather than drawing wide-reaching conclusions. Furthermore, the applicants had to use a network of pre-existing legislation and law, including the Copyright Designs and Patents Act 1988 and the Trade Marks Act 1994. This ambivalence is revealing of the UK’s current lack of comprehensive statutory guidance in relation to genAI, as well as debate within the parliament about how AI should be regulated (7). 

This case is useful in exploring the legal interactions between companies with relation to genAI but reveals little about how how existing legislation is applied to genAI might impact the day-to-day life of the British citizen. At present, the AI debate is commonly relegated to a battle between disgruntled rights-holders (artists, musicians etc…) and tech conglomerates. I argue that this is a misrepresentation of the harm, and the good that AI could do. AI has already been adopted by companies in hiring departments to streamline and filter applications and is used by many jobseekers to reduce the time it takes to apply for a position. Similarly in the healthcare context, AI may be used by radiologists to quickly and accurately review diagnoeses (8), relieving pressure on an overburdened sector. Furthermore, there’s an argument that AI’s propensity for streamlining bureaucratic tasks frees up time for humans to indulge in more creative or strategic endeavours. In these areas, it seems both responsible and natural that legislation protecting the effective usage of AI should be encouraged. 

The UK Government introduced the Online Safety Act (2023) (OSA 2023) which aims to protect the public – and particularly young people, from violent, pornographic and misinformative sources. However, this Act offers little protection from harmful stimuli introduced by AI, with experts like Prof. Lorna Woods a University of Essex internet law professor, suggesting “the law is clear but doesn’t match the market”. (9) Young people appear disproportionally at risk from harm by AI, particularly chatbots. In numerous cases, platforms from ChatGPT to Character.ai have been found to encourage or support self-harm and suicide in distressed people, and create echo-chambers encouraging the individual to isolate from friends and family. Whilst these examples may seem extreme, a YouGov survey found that 31% of 18-24 year-olds were more comfortable discussing mental health issues with an AI rather than a human therapist. This arguably stems from a core principle of current AI, pleasing the user to ensure they continue to use it. (10)

I find this last point very worrying when considering the psychological development of future generations. AI became mainstream during my first year as an undergraduate at university, and was seen largely as something cool, but not very useful given its inaccuracies and relative simplicity. As AI progressed however, and became more accurate, more people around me used it as an informative source for research, news and personal advice, and were less likely to critically evaluate its responses. Critical thinking is not something that a person is born with, it is developed and taught over years and can be enhanced or limited by a variety of factors. AI, with its constant agreement and biased responses, poses a threat to its development. It’s unlikely that there will be a law that makes critical thinking mandatory, it would seem paradoxical to do so, but perhaps there needs to be a revision of the OSA 2023 to accommodate for the very real danger that AI poses to human development. 

Harry 

Harry is an Open University student on the graduate law degree. He previously studied International Relations. Having lived in Brussels, Belgium during his teenage years, he pursued studies in terrorism and small state usage of soft power. After his law degree, he hopes to pursue a career as a solicitor. Outside of his studies, he enjoys writing, watersports, and travelling.

References

1. Japan’s emerging framework for responsible AI: legislation, guidelines and guidance https://www.ibanet.org/japan-emerging-framework-ai-legislation-guidelines
2. Singapore's AI governance regulations: Requirements and governance insights https://www.diligent.com/resources/blog/singapore-ai-regulation 
3. Artificial Intelligence Welcome to a transformative era in Artificial Intelligence in the UK. https://www.business.gov.uk/campaign/grow-your-business-in-the-uk/artificial-intelligence
4. Artificial Intelligence Playbook for the UK Government https://assets.publishing.service.gov.uk/media/67aca2f7e400ae62338324bd/AI_Playbook_for_the_UK_Government__12_02_.pdf
5. The UK’s First Copyright vs. AI Decision: Key Takeaways on a Win for the AI industry https://www.sidley.com/en/insights/newsupdates/2025/11/uk-first-copyright-vs-ai-decision-key-takeaways-on-a-win-for-the-ai-industry 
6. Stability AI largely wins UK court battle against Getty Images over copyright and trademark https://apnews.com/article/getty-stability-ai-image-copyright-trademark-fa2c561a33c7b6714a7657255a3fbdf1 
7. AI and copyright: a post-Data Bill UK timeline into 2026 https://www.pinsentmasons.com/out-law/analysis/ai-and-copyright-post-data-bill-uk-timeline-2026 
8. Redefining Radiology: A Review of Artificial Intelligence Integration in Medical Imaging https://pmc.ncbi.nlm.nih.gov/articles/PMC10487271/
9. ‘A predator in your home’ Mothers say chatbots encouraged their sons to kill themselves https://www.bbc.co.uk/news/articles/ce3xgwyywe4o 
10. AI Wants to Make You Happy. Even If It Has to Bend the Truth https://www.cnet.com/tech/services-and-software/ai-wants-to-make-you-happy-even-if-it-has-to-bend-the-truth/