🔗 Share this article

A AI company headquartered in London has prevailed in a significant high court case that addressed the legality of machine learning systems using extensive amounts of copyrighted material without permission.

Court Decision on AI Training and Intellectual Property

Stability AI, whose leadership includes Oscar-winning filmmaker James Cameron, successfully defended against allegations from the photo agency that it had violated the international photo agency's copyright.

Legal experts view this ruling as a setback to rights holders' exclusive right to profit from their artistic output, with one prominent lawyer warning that it indicates "the UK's secondary IP regime is not sufficiently strong to protect its creators."

Findings and Trademark Issues

Judicial documentation showed that the agency's images were indeed used to train the company's AI model, which allows users to generate images through text instructions. Nonetheless, Stability was also determined to have infringed Getty's trademarks in certain cases.

The judge, Mrs Justice Joanna Smith, remarked that establishing where to find the equilibrium between the interests of the creative sectors and the artificial intelligence sector was "of very real societal concern."

Judicial Challenges and Withdrawn Allegations

The photo agency had initially filed suit against Stability AI for infringement of its IP, claiming the AI firm was "completely indifferent to what they fed into the training data" and had scraped and copied millions of its photographs.

Nevertheless, the agency had to drop its initial copyright claim as there was no evidence that the training occurred within the United Kingdom. Alternatively, it proceeded with its legal action claiming that Stability was still using copies of its visual assets within its platform, which it called the "lifeblood" of its business.

Technical Intricacy and Judicial Analysis

Demonstrating the complexity of AI copyright disputes, the agency fundamentally contended that the firm's visual creation model, called Stable Diffusion, amounted to an violating copy because its development would have represented copyright violation had it been conducted in the UK.

Mrs Justice Smith determined: "A machine learning system such as Stable Diffusion which does not store or reproduce any copyright works (and has not done so) is not an 'infringing reproduction'." She elected not to make a determination on the passing off allegation and found in favor of certain of Getty's arguments about brand infringement involving digital marks.

Industry Reactions and Future Consequences

Through a statement, the photo agency stated: "We continue to be deeply worried that even well-resourced organizations such as Getty Images face substantial challenges in safeguarding their artistic works given the lack of disclosure standards. Our company committed millions of pounds to achieve this stage with only a single provider that we need proceed to address in a different forum."

"We encourage governments, including the UK, to establish more robust disclosure rules, which are essential to prevent costly court proceedings and to allow artists to protect their rights."

The general counsel for Stability AI commented: "Our company is satisfied with the judicial ruling on the outstanding claims in this proceeding. Getty's choice to voluntarily withdraw the majority of its copyright claims at the end of court testimony left only a limited number of allegations before the court, and this concluding decision eventually addresses the copyright issues that were the core issue. Our company is grateful for the attention and effort the court has put forth to settle the significant issues in this proceeding."

Wider Industry and Government Background

This judgment comes during an ongoing debate over how the current government should regulate on the issue of copyright and AI, with creators and authors including several well-known individuals advocating for enhanced safeguards. At the same time, tech companies are calling for broad access to copyrighted content to enable them to build the most advanced and efficient AI creation platforms.

Authorities are presently consulting on IP and AI and have stated: "Lack of clarity over how our intellectual property system operates is holding back development for our artificial intelligence and creative industries. That must not persist."

Legal experts following the situation suggest that regulators are examining whether to implement a "text and data mining exemption" into UK IP legislation, which would permit copyrighted works to be used to train machine learning systems in the UK unless the rights holder opts their works out of such development.