Getty vs Stability AI Ruling Leaves AI Copyright Law in Legal Limbo

Court Delivers Mixed Verdict in Landmark Case

The English High Court has delivered its highly anticipated ruling in Getty Images v Stability AI, a case that many hoped would provide definitive guidance on the intersection of artificial intelligence and copyright law. However, the 205-page judgment handed down by Mrs Justice Joanna Smith in late November has left the legal landscape as murky as before, with both sides claiming partial victory in what has become a pyrrhic legal battle 1

1

Getty Images initiated legal proceedings in January 2023, becoming the first major media company to challenge an AI firm in UK courts over alleged unauthorized use of copyrighted content. The case centered on allegations that Stability AI's Stable Diffusion model, which generates AI images from user prompts, had been trained on approximately 12.3 million visual assets from Getty Images and other publicly accessible websites without permission 1

1

Source: Creative Bloq

Jurisdictional Limitations Undermine Primary Claims

The most significant aspect of the ruling was what it failed to address rather than what it resolved. Getty's primary copyright infringement claim was dismissed not on its merits, but due to jurisdictional constraints. The court found that the alleged training of Stable Diffusion occurred on servers outside the UK, placing it beyond the reach of the Copyright, Designs and Patents Act 1988 (CDPA) 2

2

This jurisdictional limitation meant that the court never addressed the fundamental question that the creative industry was hoping to see resolved: whether training a generative AI model on copyrighted material within the UK, without authorization, constitutes copyright infringement. As Justice Smith acknowledged, "The findings are both historic and extremely limited in scope" 1

1

Secondary Infringement Claims Fall Short

With the primary claim dismissed, the case hinged on allegations of secondary copyright infringement. Getty argued that Stability AI had imported an "article" - the Stable Diffusion model itself - knowing it contained infringing copies of Getty's images. However, the court ruled that while an "article" can include electronic copies stored in intangible form, Stable Diffusion's model weights never actually contained or stored copies of the Getty images 2

2

The ruling established that for something to qualify as an infringing copy, there must be an explicit reproduction of the copyrighted work. Model weights, which represent the way an AI model learns during training, do not contain visual information from copyrighted works in a form that constitutes reproduction 1

1

Limited Victory on Trademark Grounds

Getty did achieve a partial victory on trademark infringement claims. The court found that earlier versions of Stable Diffusion had generated images bearing Getty Images and iStock watermarks, constituting trademark infringement. However, this success was limited in scope, with damages likely to be nominal 2

2

The trademark finding does underscore the need for AI developers to implement appropriate safeguards to minimize the risk of reproducing protected trademarks in AI-generated content, but it falls far short of the comprehensive legal framework that content creators had hoped the case would establish.

Related Stories

UK Court Delivers Mixed Verdict in Getty vs. Stability AI Copyright Case, Leaving Key Questions Unanswered

Getty Images Drops Key Copyright Claims Against Stability AI in Landmark UK Lawsuit

Getty Images and Stability AI Face Off in Landmark UK Copyright Trial

U.S. Judge Allows Artists' Copyright Lawsuit Against AI Companies to Proceed

Implications for the AI Industry and Content Creators

For AI developers, the ruling provides some temporary relief, particularly for those training models outside the UK. However, legal experts warn that this apparent victory may be superficial. As copyright expert Cerys Wyn Davies noted, the decision highlights how AI developers can potentially circumvent UK copyright protection by ensuring that copying for training purposes takes place outside UK jurisdiction 1

1

For content creators and rights holders, the decision represents a frustrating setback. The dismissal of copyright claims underscores the constraints of the UK's existing intellectual property framework when applied to generative AI technologies. Without statutory reform, enforcing rights over copyrighted data used to train AI models will remain challenging, particularly where training occurs outside the UK 2

2

Regulatory Gaps Exposed

The judgment has exposed significant gaps in existing UK intellectual property law, particularly regarding territorial reach and jurisdictional boundaries for globally trained AI systems. In response to the ruling, Getty called for stronger transparency measures to reduce costly disputes and better safeguard creators' rights 2

2

The UK Government now faces mounting pressure to strengthen transparency obligations for large-scale AI development and clarify jurisdictional boundaries. Ongoing consultations with expert groups from both creative and technology sectors will be closely watched as policymakers attempt to balance the protection of human creativity against the promotion of AI innovation 2

2