An argument between Getty Images and Stability AI about some weird photographs and wonky watermarks has turned into the most significant copyright case in recent history.
It will decide just how powerful UK copyright really is. It comes at an interesting time, putting the focus on the High Court as it grapples with these issues whilst the House of Commons succeeds in its play for more time.
Quick re-cap
Getty’s case is that Stable Diffusion was trained on vast quantities of Getty-licensed images without permission, and can reproduce them (including, on occasion, the Getty trademark) on request.
Under UK law, that raises copyright, trademark and database right infringement issues.
What has happened so far
Despite being the first UK case to address these critical issues, in its first week the trial was almost de-railed twice because of a niche point around the wording of the pleadings.
The judge is resistant to adjourn the trial, given it has cost both parties millions and its outcome is so hotly anticipated, but the possibility of a delay remains live. The impressive spectre cast by a “nice-to-have” point (in the judge’s words) shows the monumental importance of this case to both parties.
Why this matters for UK tech
But it is not just the parties who will be impacted – the outcome of this case will be critical in shaping the interaction between the UK’s tech and creative industries.
The UK government has made no secret of its ambition to be a home for AI development, but the Getty case illustrates how far our current statutory framework lags behind the technology.
AI models trained abroad, using data scraped without permission, are being deployed in the UK without clear answers as to whether this is legal.
For UK AI developers, the trial offers a sneak peek into how courts could interpret “use” in a training context, how they will balance open justice against confidentiality in model architecture, and how much protection rightsholders can actually expect from current UK law.
A finding in Getty’s favour could mean uncomfortable levels of exposure for models trained without the right licensing arrangements in place, and companies whose AI offerings sit on top of those models.
Downstream, it could lead to a reduction in tech offerings available to UK consumers, as developers prioritise jurisdictions with more relaxed rules.
A decision favouring Stability AI is preferable for UK tech companies but could equally lead to disruption. The prospect of certain AI training practices or locations being outside the scope of UK copyright has alarmed rights holders. A judicial decision confirming this would almost certainly trigger demands for faster legislative intervention.
While a final decision is unlikely before late 2025, the case already offers useful guidance. In this author’s view, current legislation requires companies offering AI models to UK consumers to comply with UK copyright law.
If the court comes to the same conclusion, those requirements will apply immediately and a “bill of clean training” could be invaluable – at least until the government takes stock and finally makes its legislative move.
Rebecca Newman is managing associate in Addleshaw Goddard’s commercial disputes team
The post Getty v Stability AI: The courts step in while the Commons step back appeared first on UKTN.
