The tension between the opportunities offered by new technologies and the need for artists to be able to control the use of their own works and derive revenue from them is all too familiar. Inevitably, case law and/or legislation will draw an artificial line between what is fair and what is not.
In recent months, a number of court cases have been filed regarding the use of images of artwork by tech companies to “train” their artificial intelligence (AI) tools. These companies grab images from the internet and use them to program their AI with different themes, moods, and styles. The primary target of these lawsuits is Stable Diffusion, Stability AI’s image-generating tool. Getty Images is suing Stability in the US and UK for allegedly using millions of images from Getty’s library to form Stable Diffusion. In the United States, Getty would claim damages of 2,000 billion dollars!
The issues with AI and copyright law are manifold: first and foremost, the question of how and from what the AI tool learns from, raises questions as to whether the learning process (regardless of the outcome) infringes copyright. The cases against Stability are the simplest example: Getty’s allegation is that their copyrighted images were simply copied during the process of forming the Stable Diffusion tool, and that copy (in assuming it was unlicensed) would infringe any existing copyright in the images. It’s hard to see how Stability can defend this, as any programming process will involve reproducing the source material, even if the copy is only stored very briefly. In the US, “fair use” issues may well be relevant to the stability case, but this is less applicable in the UK.
It’s a much more nuanced question of whether the results produced by AI tools are themselves infringing works. In practice, this will likely involve an assessment to determine whether part of an original image was copied and whether that part was “substantial”. Some claim that Stable Diffusion was used to generate images in the style of named artists (even though their actual work was not reproduced), further complicating the analysis of infringement.
Stability cases will be carefully monitored to achieve some judicial certainty on these issues. But, with appeals, it could be many years before courts catch up with today’s technology, let alone tomorrow’s.
It’s a familiar tension between the opportunities offered by new technologies and the need for artists to be able to control the use of their work and derive income from it. Inevitably, case law and/or legislation will draw an artificial line between what is fair and what is not.
Steps are underway to legislate on the issue of counterfeiting. In the UK, the government has recently proposed to extend an exception to the infringement rules that currently exist for data mining for non-commercial research purposes, to allow this for any purpose, allowing the training of AI tools without infringing. , but the EU is moving forward with a similar exception that would apply unless the rights holder has expressly reserved their rights, a compromise solution that raises even more problems for both sides in the debate.
The other side of the coin is the protection of the AI-generated works themselves. Can works generated by AI be “original” in order to be themselves covered by copyright? In the context of an AI tool that learns styles and images, this can be a very difficult question to answer and again may involve very granular case-by-case analysis.
And if copyright exists in the created art, who owns it? UK copyright law states that the author or creator of an original work owns the copyright. That author then has the right to reproduce that work or permit others to do so under a full assignment of copyright or license. However, what if the digital asset is not produced by a human but was produced by artificial intelligence? Who owns the copyright in something created by a machine where there is no “artistic effort” or “labor, skill and judgment” of the artist? A good example are the CryptoPunks characters, which are randomly generated computer images, each different from the previous one, but all with a common 8-bit format. Who owns the copyright to these computer-generated images?
In the UK, the Copyright, Designs and Patents Act 1988 is unusual in its attempt to address this issue as “computer-generated works”. It defines the owner (or author) as the person who makes the “arrangements necessary for the creation of the work”, but this will not always be so easy to determine, and it can be difficult to determine where and what elements humans has “made arrangements”.
Stability cases will be carefully monitored to achieve some judicial certainty on these issues. But, with appeals, it could be many years before courts catch up with today’s technology, let alone tomorrow’s.
- Hetty Gleave, Partner and Art Law Specialist, and Eddie Powell, Partner and Intellectual Property and Copyright Specialist, at Fladgate LLP
