Artificial Intelligence and Intellectual Property: an update - Law Firm - Fox Williams

As artificial intelligence (AI) continues to develop rapidly, intellectual property (IP) law is having to try to keep up. This can be seen in several recent and ongoing developments in the UK.

Two recent cases and a consultation examining the intersection of AI and IP

1. From 9 June 2025 to 30 June 2025, the English High Court heard the much-anticipated case of Getty Images v Stability AI. Getty Images alleged several forms of IP infringement, including:

Copyright and/or database rights infringement – Getty Images claimed its images were unlawfully used to train and develop Stability AI’s model. This claim was later withdrawn – which is disappointing in terms of obtaining further court guidance – potentially as the alleged training/development may have occurred outside the UK jurisdiction.

Secondary copyright infringement – Getty Images argued that Stability AI was importing the trained AI software (arguing this fits within the meaning of an “article”) into the UK and knows or has reason to believe it is an infringing copy.

Trade mark infringement and passing off – Getty Images claimed that AI-generated outputs unlawfully reproduced its “gettyimages” watermark.

We await the judgment which will hopefully provide further guidance on AI training and AI inputs and outputs and whether these can fall within existing IP infringement rules.

2. On 21 and 22 July 2025, the Supreme Court (the UK’s highest court) heard an appeal by Emotional Perception AI. The question at the heart of the case: can an artificial neural network – a type of AI machine learning model inspired by the structure of the human brain – be eligible for patent protection. We covered the previous Court of Appeal case here and we await the Supreme Court judgment.

3. The UK government’s public consultation on AI and copyright closed at the end of February 2025 and attracted over 11,500 responses – highlighting the complexity and significance of this issue. On 16 July 2025, the government announced a new expert working group of representatives from the creative industries and the AI sectors. This is to help balance protecting the rights of creatives with continued innovation in the AI sector (see here).

In the meantime, key uncertainties remain – particularly for businesses that develop AI tools, integrate them into products or rely on third-party AI models and datasets. We explore below examples of such grey areas in IP:

A. Copyright – Copyright may subsist in, for example, sections of text, code, images, photographs, videos and music. However, it can be unclear how AI works are protected by copyright (e.g. can AI generated works meet the “originality” requirement if they are based on common prompts or on pre-existing works?), who owns the copyright in AI generated content (e.g. the user providing the input to generate the content or the AI provider?), can AI developers lawfully use copyrighted works to train their AI and are there any copyright exceptions that could apply (e.g. the UK is considering extending the exception for text and data mining so it is not limited to non-commercial purposes), who is liable if an AI infringes copyright (e.g. the user who entered the prompt into the AI or the company that trained the AI?) and when does UK copyright law apply (e.g. what if the AI model is trained abroad but used in the UK)? See also our previous article on AI and copyright here.

B. Patents – These can protect novel inventions. There are similar questions to many of those in the copyright section above, but we have more certainty in that the UK Supreme Court has held that an AI could not be named as an inventor (see our article here). As explained at the start of this article, whether ANNs can be protected by patents is to be decided by the Supreme Court.

C. Database rights – Database rights protect the investment in collecting the data and creating the database. This could be relevant, for example, when using structured data sets of third parties. There are similar questions to many of those for copyright, including whether databases created by or using AI can be protected by database rights (e.g. could they meet the required substantial investment in obtaining, verifying or presenting the contents of the database?), who owns the database rights in AI generated content, and whether an AI taking data from a protected database could be extracting or re-utilising a substantial part of a database to constitute infringement.

D. Trade marks (and passing off) – Key questions include where trade marks or protected unregistered marks appear in prompts, training material or outputs and/or are used in inputs, could this be trade mark infringement or passing off, and are there any exceptions that could apply (e.g. is this trade mark use and could some uses fall within the trade mark defence for honest descriptive use?). We may get more guidance from the Getty case.

E. Confidential information – If data is confidential or is protected by a non-disclosure or confidentiality agreement, what if this was put into an AI and the AI was trained on confidential information? Would this be a breach of your confidential information?

Advice for AI companies and startups

As the legal landscape continues to evolve, so too does the scrutiny on how AI companies build and scale their models. These aren’t just technical decisions – they’re legal and commercial ones too. With the current uncertainty over AI and IP law, it is advisable to ensure that you have terms in place clarifying the position. If you are unsure whether your training data, model architecture, outputs or your usage of AI systems could infringe third-party rights, then please feel free to reach out.

We advise businesses (including those in the technology, fashion and publishing sectors) coming to grips with the new world of AI and how it can assist their businesses to grow. We also advise AI startups across the stack – from pre-seed to post-exit – on how to manage IP risk without stifling innovation.