🔗 Share this article

A artificial intelligence company based in London has won in a significant high court case that examined the lawfulness of AI models utilizing extensive quantities of protected data without authorization.

Judicial Ruling on AI Training and Copyright

The AI company, whose directors includes Oscar-winning director James Cameron, effectively defended against claims from Getty Images that it had violated the global photo company's copyright.

Industry observers consider this ruling as a setback to rights holders' sole ability to profit from their creative output, with a senior attorney cautioning that it indicates "the UK's secondary copyright regime is not sufficiently robust to safeguard its creators."

Evidence and Brand Issues

Court evidence showed that the agency's photographs were indeed used to develop Stability's system, which allows users to generate images through text instructions. Nonetheless, Stability was also found to have infringed the agency's brand marks in certain cases.

The judge, Mrs Justice Joanna Smith, stated that establishing where to find the balance between the concerns of the artistic sectors and the AI sector was "of significant public importance."

Legal Complexities and Withdrawn Allegations

Getty Images had originally sued Stability AI for infringement of its intellectual property, claiming the AI firm was "entirely unconcerned to what they fed into the training data" and had scraped and copied millions of its images.

However, the company had to drop its initial IP case as there was insufficient evidence that the training occurred within the UK. Alternatively, it continued with its legal action arguing that Stability was still using reproductions of its image content within its systems, which it called the "lifeblood" of its business.

System Intricacy and Judicial Reasoning

Highlighting the complexity of AI copyright cases, the agency fundamentally contended that Stability's visual creation model, known as Stable Diffusion, amounted to an infringing reproduction because its creation would have constituted copyright infringement had it been carried out in the United Kingdom.

Mrs Justice Smith ruled: "A machine learning system such as Stable Diffusion which does not store or replicate any copyright material (and has never done) is not an 'violating copy'." She elected not to rule on the misrepresentation claim and found in support of certain of the agency's arguments about trademark infringement related to digital marks.

Sector Responses and Ongoing Implications

Through a official comment, Getty Images stated: "We remain profoundly worried that even well-resourced companies such as our company encounter significant difficulties in protecting their creative output given the absence of transparency requirements. Our company committed millions of currency to achieve this point with only a single company that we need proceed to address in another forum."

"We encourage authorities, including the United Kingdom, to establish more robust disclosure regulations, which are crucial to prevent expensive legal battles and to enable creators to protect their rights."

The general counsel for Stability AI commented: "Our company is pleased with the judicial ruling on the outstanding allegations in this case. Getty's decision to voluntarily withdraw most of its IP cases at the conclusion of trial proceedings resulted in a subset of allegations before the court, and this concluding decision ultimately addresses the copyright issues that were the core matter. We are grateful for the time and effort the judiciary has put forth to settle the significant issues in this case."

Wider Sector and Regulatory Background

The ruling emerges amid an ongoing debate over how the current government should regulate on the matter of intellectual property and AI, with creators and authors including numerous well-known individuals advocating for greater protection. Meanwhile, technology companies are advocating wide access to protected material to allow them to develop the most powerful and effective AI creation systems.

The government are currently seeking input on IP and artificial intelligence and have declared: "Lack of clarity over how our intellectual property system operates is impeding development for our AI and artistic sectors. That cannot continue."

Legal experts following the situation indicate that regulators are considering whether to introduce a "text and data mining exemption" into British IP legislation, which would allow copyrighted works to be utilized to train machine learning systems in the United Kingdom unless the owner opts their works out of such training.