Getty Images (US), Inc & Ors v Stability AI Ltd

I consider that a penal notice is not necessary in the circumstances of this case. A voluntary undertaking has been given by the defendant, as is shown on the face of the order. The defendant knows perfectly well what it is undertaking to do and not to do and it is not clear to me who the third parties are that need to be given notice in relation to this. Certainly no evidence has been served in respect of those third parties.

As to the issue of enforcement in due course and the potential for a committal application, all versions of the relevant models have now been retired and it is not, in those circumstances, clear to me why the claimants would need a penal notice which appears to me to be unnecessarily inflammatory.

For those reasons, I am going to reject the inclusion of a penal notice, but as I say, if anyone wants a longer judgment, they should feel free to ask me for one.

I can deal with this briefly. I do not consider that what I have said in the judgment binds me as to how I approach the question of whether to dismiss or discontinue the claims of primary infringement of copyright and database right infringement.

In my judgment, in circumstances where those claims were pushed all the way to closing submissions, at vast expense and following substantial cross-examination, and in circumstances where it was then accepted by the defendants in closing that they could not pursue those claims in light of the evidence (including that they said in a letter of 20th June 2025 that their decision to abandon those claims “reflects the impact of oral evidence” on all the claims) the appropriate order is that those claims should be dismissed.

However, it is also, in my judgment, important that the circumstances in which those claims have been dismissed should be clear on the face of today’s order. Accordingly, either in the recitals to the order or in paragraph 5, there should be added words to the effect of: “... in circumstances where those claims were fought all the way to closing submissions and then abandoned”. This will ensure that there can be no later suggestion (along the lines that Mr. Marshall has identified), whether in the UK or elsewhere, that the court has actually adjudicated on those claims.

I must now deal with an application by the claimants for permission to appeal in relation to my dismissal of Secondary Infringement of Copyright Claim.

I am going to grant permission to appeal in relation to this issue. The appeal does, in my judgment, have a real prospect of success. It concerns a pure question of law, namely a matter of statutory construction on which the minds of reasonable lawyers may differ.

Furthermore, as Miss Lane rightly says, there are other compelling reasons for the appeal. The point of law is both novel and important because it concerns how the provisions of the CDPA should be construed (and specifically the phrase “infringing copy”) in the context of an AI model. This is not an issue that has previously been considered by any court. As the claimants submit, that novel point also has potentially far-reaching ramifications for AI models and intangible articles such as software more generally.

In my judgment this is plainly an issue that would benefit from consideration by the Court of Appeal. That does not necessarily mean that I accept that every single one of the proposed grounds of appeal has a real prospect of success, but I see no reason to salami-slice them now. Much better that the claimants should have the opportunity to raise on appeal all of the points they wish to raise in this context.

For those reasons I will grant permission in relation to all of the grounds of appeal.

I am going to refuse permission to appeal in relation to the defendant’s application to appeal my findings of trade mark infringement. In my judgment, an appeal on the three main grounds that have been identified has no real prospect of success.

Grounds 1 and 2 are plainly an attempt to appeal my factual findings – essentially the defendant says that I could not have arrived at the decisions I made as to UK activity on the basis of the available evidence. However, I reviewed in detail the available evidence over many paragraphs of the judgment (including recording relevant concessions made by the defendant) and it was neither wrong nor rationally insupportable to make the findings I made on that evidence. I do not consider there to be any prospect of overturning those findings.

In relation to Ground 3, the focus is on my identification of the average consumer in the particular context of watermarks. I refuse permission in respect of this ground also, essentially for the reasons that Ms. Bowhill identified during her helpful submissions. As to the specific subparagraphs in Ground 3: at sub-para (ii), the defendant suggests that my findings ignore the twofold purpose of watermarks (to render an image unusable and to identify the owner of the right for licensing purposes) and that in holding (correctly) that the vast majority of users presented with the sign would discard it, I failed to recognise that such consumers would not be approaching the sign as a consumer and did not accord it the attention of the average consumer. There is no prospect of success on this ground. There was no evidence at trial to support this proposition and it is not explained by the defendant why the consumer could not be responding to both purposes of the trademark at the same time. At sub-paragraph (iii) it is suggested that an AI user would pay a high level of attention to the sign. But this is consistent with a finding in my judgment at [284] that some average consumers would pay a high degree of attention and I cannot see where it takes the defendant. At sub-paragraph (iv) an argument is raised as to the illogicality of a consumer supposing that a sign was a communication of trade origin. However, this attributes a level of sophistication and analysis to the average consumer which goes beyond what is appropriate and in respect of which there was no supportive evidence at all. At sub-paragraph (v) the defendant suggests that I was wrong to rely upon a particular SalesForce message, suggesting that that message was clearly capable of different interpretations. However, this is no more than a submission that another judge might have taken a different view – which is the wrong test. I reject the suggestion that the defendant has a real prospect of establishing that my interpretation of this message was one that no reasonable Judge could have arrived at.