Next week marks the commencement of hugely significant legal proceedings between Getty Images and Others and Stability AI.
This highly anticipated trial, set to begin on 9 June 2025, will last three weeks in the UK High Court. The case is poised to address critical issues at the intersection of copyright law and artificial intelligence technology. With concurrent litigation taking place in the United States, the outcome of this trial is expected to have far-reaching implications for the legal landscape surrounding AI and intellectual property rights. It will likely result in one of the first substantive judgments in the spate of disputes taking place worldwide, particularly in the US.
Background to the Case
The claimant, Getty Images (Getty), is a visual media company and a leading licensor of stock photography images. Stability AI (Stability) defending the claim is a London-based developer of generative AI systems, including its 'Stable Diffusion' models which generate images from text or image prompts input by users of the models.
On 16 January 2023, Getty filed a claim against Stability in the UK High Court, asserting inter alia that the company scraped and used millions of Getty's copyrighted works without consent to train its AI model. Getty also contends that the output produced by Stability's Stable Diffusion models in response to user prompts infringes Getty's copyright by reproducing substantial parts of the works used to train the model.
Key issues
There are understood to be over 60 issues for the High Court to grapple with during the 3-week long trial, but the headline points of contention are as follows:
1. Copyright Infringement
Getty alleges that Stability has infringed copyright in millions of works in respect of which Getty is either the copyright owner or the exclusive licensee (the Getty Content) by (a) using the Getty Content without Getty's permission in order to train its generative AI models; and (b) by reproducing substantial parts of the Getty Content in the output generated by the Stable Diffusion models.
The Training and Development Claim
Getty claims that Stability trained its Stable Diffusion models on the LAION-5B dataset which Getty states comprises 5.85 billion image-text pairs which were created by scraping links to photographs and videos from various websites, including Getty's websites. Accordingly, Getty claims that Stability trained Stable Diffusion on the Getty Content. Getty relies on evidence of Stability having several UK-based employees and the fact that Stability's website is in the English language and features a Union Jack on the home page to demonstrate that Stability's Dream Studio platform (which enables Stability's customers to access or download the Stable Diffusion model) targets UK users.
Getty claims that Stability used the scraped links in the LAION-5B dataset to revisit Getty's websites and copy some of the Getty Content featuring therein. Getty claims that such content was downloaded and stored by Stability in order for them to add "noise" to it so as to alter the content and make it harder to discern what is visually represented because it has been cluttered by other elements that are not part of the original image. Getty claims that this process involves the creation and storing of further copies of the Getty Content. Getty claims that further copies are created and stored by Stability when it decodes the Getty Content with added "noise" in order to train Stability Diffusion. Getty claims that each of these acts of copying without Getty's consent amounts to infringement of Getty's exclusive right to reproduce the Getty Content (under s.17 of the CDPA).
Stability has acknowledged that some images from Getty's website were used during the training of Stable Diffusion, but denies the claims of copyright infringement. Stability contends that the development and training of its Stable Diffusion models occurred outside of the UK, and therefore, no acts restricted by Getty's intellectual property rights were performed in the UK.
The Output Claim
Getty also claims that the output generated by Stable Diffusion in response to user text and image prompts reproduces a substantial part of the Getty Content, thereby infringing Getty's exclusive rights to reproduce the same and that it infringes Getty's exclusive right to communicate the Getty Content to the public (under s.20 of the CDPA) and to authorise such acts of reproduction and / or communication to the public (s.16(2) CDPA).
Getty claims that in some cases, the output generated by Stable Diffusion in response to user text prompts bears the GETTY IMAGES and ISTOCK trade marks as a watermark. In addition, Getty claims that where users of Stable Diffusion enter image prompts (i.e. where users enter an image into Stable Diffusion in order to produce a new image) which comprise the Getty Content, that the output generated by Stable Diffusion in response reproduces the whole or a substantial part of the Getty Content.
Getty also alleges that Stability is liable for importing an "article", namely Stable Diffusion, into the UK which Getty knows or has reason to believe is an infringing copy of the Getty Content contrary to s.22 of the CDPA; and that Stability is guilty of possessing in the course of a business, selling or letting for hire, or offering or exposing for sale or hire, an "article", namely Stable Diffusion, which is and which Stability knows or has reason to believe is, an infringing copy of the Getty Content contrary to s.23 of the CDPA.
Stability contends that its models' synthetic image outputs in response to user prompts are created without use of any of Getty's copyright-protected works, that the outputs do not reproduce any copyright works and that any resemblance between generated images and Getty's works is coincidental and not due to copying. Stability also contends that all output generated by the Stability modes is a result of the actions of users of the models, not Stability, and that Stability has not authorised any infringements and has taken reasonable steps to prevent any such infringements.
2. Trade mark Infringement and passing off
Getty claims to have several UK registered trade marks comprising the words "GETTY IMAGES" and "ISTOCK" in several classes and claims to have goodwill and reputation in these marks, as evidenced by the number of UK-based customers that conduct searches on Getty's websites each year and Getty's advertising of its services in the UK. Getty states that each photograph and image that features on Getty's website displays a watermark that contains one or other of Getty's registered trade marks.
Getty accuses Stability of trade mark infringement and deceptive trade practices that take unfair advantage of Getty's rights in its trade marks. Getty claims that Stability used, without authority, Getty's trade marks (through the watermarks in the Getty Content) in order to train Stable Diffusion and that they are used in the images generated in response to user prompts by the Stable Diffusion models which include the marks as watermarks.
Stability denies using any of Getty's trade marks in the course of trade. It argues that any watermarks appearing in generated output are not used by Stability and are not an indicator of trade mark use.
3. Database Right Infringement
Getty also alleges database right infringement on the basis that Stability has, without authority, extracted and re-utilised a substantial part of the content in Getty's database in order to train the Stable Diffusion models. In its particulars, Getty provided detail on the extent of the monetary investment it has made in obtaining, verifying and presenting the contents of its database. However, Stability challenges Getty's assertion of database rights, arguing that Getty has not adequately demonstrated the subsistence of such rights in its particulars in order to allege infringement of the same.
Decisions so Far
Since the claim was filed by Getty in January 2023, there have been interlocutory applications dealing with various issues that will be addressed further at trial. We have summarised below the outcomes of the relevant applications.
1 December 2023 Judgment - Application for Summary Judgment
In late 2023, both parties made applications to the High Court. Stability AI applied for reverse summary judgment in respect of the following claims set out in Getty's Particulars of Claim on the ground that Getty has no real prospect of succeeding at trial on these points, namely: (1) the claim that Stability infringed Getty's copyright and database rights by downloading on servers and/ or computers in the UK visual assets and associated captions in order to develop and train Stable Diffusion (the Training and Development Claim); (2) the claim of secondary infringement of copyright by Stability's importation of an "article", being its pre-trained Stable Diffusion software, into the UK (the Secondary Infringement Claim); and (3) the claim of trade mark infringement and passing off relating to the output images generated by Stable Diffusion (the Trade Mark and Passing Off Claims). Following an amendment made by Getty to its particulars providing that they no longer intended to rely on the case of Louboutin v Amazon (C-148/21 and C-184/21, EU:C:2022:1016), Stability dropped the Trade Mark and Passing Off Claims from its application for reverse summary judgment.
After reviewing the evidence presented by the parties, including extensive evidence from Stability arguing that all training and development of Stable Diffusion took place outside of the UK, High Court judge Mrs Justice Joanna Smith concluded that the Training and Development Claim has a real prospect of success at trial because (a) Stability has its principal place of business in the UK and has a number of UK-based employees in its development team and there was sufficient evidence adduced by Getty to necessitate fuller investigation into the facts relating to this issue; and (b) Getty's evidence as to the computer resources used by Stability in the UK raises legitimate factual questions which cannot be resolved without further investigation at trial. Mrs Justice Joanna Smith also considered there were several conflicts of fact on the evidence which made the claim unsuitable for summary judgment. Such conflicts included: when work on Stable Diffusion commenced; the roles played by Stability's witnesses in the development of Stable Diffusion; and what computing resources were used to train and develop Stable Diffusion.
The Secondary Infringement Claim centres around the definition of an "article" and whether this could include an intangible thing, such as the Stable Diffusion models. Mrs Justice Joanna Smith was not convinced on the evidence adduced by Stability that Getty is bound to fail in arguing that an "article" could include intangible things. Accordingly the summary judgment application was also dismissed in relation to the Secondary Infringement Claim on the basis that this is a point of law which must be decided by the trial judge.
Stability also made a request for further information with a view to obtaining examples of the copyright works alleged to have been infringed. This issue was resolved at the hearing, with Getty agreeing to provide examples of the copyright works it alleged had been infringed.
Getty applied to the Court for permission to amend its particulars to include an additional claim relating to an image-to-image feature of Stable Diffusion which Getty argues permits users of Stable Diffusion to make "essentially identical copies of copyright works" (the Output Claim). This amendment was opposed by Stability on the basis that Getty has no prospect of success and that the claim is inadequately particularised. Stability argued that when a user downloads a copy of Stable Diffusion and enters an image prompt in order to generate a new image that may amount to a copy of a copyright work, it is the user that brings about the copying and not Stability, which merely acts as an intermediary or host allowing access to its platform but that there is no act of copying by Stability. Whereas, Getty claims that while the original image is obtained by the user and entered into the Stability Diffusion software as a prompt, the synthetic output generated by the software is the result of differing levels of corruption by noise which has been authorised by Stability and communicated to the public by them. Based on these submissions, Mrs Justice Joanna Smith was convinced that Getty had a real prospect of success on this claim and that, on balance, Getty's claim should go to trial. Accordingly, Getty was permitted to add this claim to its particulars.
14 January 2025 Judgment
Getty's case included a representative action on behalf of around 50,000 photographers and content creators whose works were licensed exclusively to Getty. Getty claimed that these individuals were represented by the Sixth Claimant, an company by the name of Thomas M Barwick Inc. The company's director, Mr Thomas M Barwick, is a prominent photographer. Getty's case was that this is the number of photographers that are the owners of copyright subsisting in content licensed on an exclusive basis to Getty and that, as a result, each of these copyright owners has a concurrent action with Getty to restrain Stability's alleged acts of copyright infringement. Getty argued that the Sixth Claimant represented the same interests in the claim as those 50,000 owners of copyright-protected works licenced on an exclusive basis to Getty.
On 14 January 2025, Mrs Justice Joanna Smith in the High Court handed down a judgment following an application by Stability which sought an order that the Sixth Claimant could not act as the representative for this group of individuals. The Judge had to determine whether the Sixth Claimant had the same interest as the represented parties (per CPR r.19.8) and whether the class could be adequately defined, as this is necessary in determining whether members of that class have the same interest. Getty's proposed class definition was as follows:
"...the class of persons represented by the Sixth Claimant are those who are owners of the copyright subsisting in artistic works and film works that have been licensed on an exclusive basis to the First Claimant, the copyright in which has been infringed by the Defendant. The persons that fall within such class can be identified on the basis that (i) they have entered into an exclusive licence with the First Claimant in respect of artistic works and/or film works; and (ii) the exclusively licensed works include works which were used to train Stable Diffusion...".
The Judge found that Getty's proposed class definition was problematic as it depended on the outcome of the litigation (i.e. whether there had been copyright infringement) and on Getty identifying which works owned by the represented class were used to train Stable Diffusion, which could not be determined by either side at the time of the hearing. Consequently, the Judge concluded that the representative action could not succeed as it was not possible to say of any particular individual in the class whether or not they qualified for membership of the represented class and the Judge could not be satisfied that the representative claim would remove the need for an expensive and time-consuming individualised assessment of the issues of liability and quantum relating to the proposed representative parties or that this would not create a significant case management burden for the Court.
Getty's alternative proposal to proceed without joinder of all exclusive licensors (per CPR r.19.3) was also rejected on the basis that they had not provided sufficient evidence to assure the Judge that there would be no prejudice to Stability from potential future claims.
However, the Judge suggested that Getty could make another application to the Court for permission under CPR 19.3 for the copyright owners to bring their own claims to allow the issues to be dealt with using sample cases. Such an application would require satisfactory evidence adduced by Getty that, in obtaining such consent, there would be no future claims arising against Stability. The Judge also suggested that Getty could narrow the scope of the represented parties to facilitate effective case management.
Looking ahead
This is a landmark case, the outcome of which is anticipated to have far-reaching implications for the use of copyrighted material in AI model training, potentially influencing how AI developers manage data and how creators protect their works from unauthorised use. No doubt the High Court's judgment will also impact the UK Government's proposed way forward, following the UK IPO's recent consultation on reforming UK copyright law in light of advancements in artificial intelligence and the UK Government's policies around incentivising innovation.
For more AI and intellectual property insights, visit our Insights page at: www.simmons-simmons.com/insights.

.jpg?crop=300,495&format=webply&auto=webp)
.jpg?crop=300,495&format=webply&auto=webp)




.jpg?crop=300,495&format=webply&auto=webp)
_11zon.jpg?crop=300,495&format=webply&auto=webp)



_11zon.jpg?crop=300,495&format=webply&auto=webp)
_11zon.jpg?crop=300,495&format=webply&auto=webp)





