There have been several high-profile intellectual property headlines over the last 12 months, both in the US and England and Wales, writes Maria Peyman, Partner with Birketts LLP.
The intellectual property right at the core of these headlines is copyright, which protects works such as novels, academic papers, works of art, sculpture, photographs, software code, music and films.
For the purposes of the UK, copyright arises automatically in original works, with the copyright belonging to the ‘author’ (except where the author created the work in the course of their employment, in which case copyright is owned by the employer).
The length of copyright protection depends on the nature of the work and whether it is computer generated but is between 50 years from creation and 70 years after the death of the author.
In addition to the headlines referred to above, copyright has been at the centre of the storm surrounding one of the most cutting edge of technologies, artificial intelligence (AI).
Since its launch, users have been keen to try out generative AI platforms such as ChatGPT, generating everything from music to images to essays. Its output has often been questioned in terms of accuracy and even the legal profession has not been immune. Both in the US and England there have been reports of parties relying on case law which was in fact invented by ChatGPT following a search.
Despite the fanfare and engagement by the public, AI programmes such as OpenAI, Stability AI and other AI platform creators such as Microsoft find themselves in the eye of an intellectual property rights storm. With many of them fielding allegations, including claims of copyright infringement from organisations such as Getty Images, the New York Times and individual authors.
In simple terms AI is ‘trained’ by using information which is (according to OpenAI) “(1) publicly available on the internet, (2) information that [OpenAI] licenses from third parties, and (3) information that [their] users or human trainers provide.”
Consideration of this statement readily identifies the potential source of the fight. Merely because content can be found online or is publicly available does not mean it is not protected by copyright.
The complaint of unauthorised use of images online is not unusual. An unwitting use of a photo or image found online can often find businesses in receipt of correspondence from the rights owner in the image complaining of copyright infringement and seeking damages. In order to use ‘publicly available content’ without risk one must obtain express permission from the copyright owner (and they may require a fee in exchange for the use). If a small organisation uses someone else’s work without permission, there may be a quick resolution between the parties and potentially a sum of damages to pay.
However, if you are an AI creator and your product has been developed by using publicly available material but without express permission or license it may not be as easy to resolve.
Many of the claims so far have been commenced in the US, including by individual authors, but there is one claim commenced in England and Wales by Getty Images against Stability AI, for alleged copyright infringement in relation to Getty’s images.
If this case gets to trial it may provide detailed insight as to how the AI was trained and how the ‘publicly available’ material was used. On a legal level there will, of course, be significant interest seeing how existing legislation fits with modern technology.
Just recently, OpenAI has stated that without access to copyright material it would not be possible to train large language models. So, any judgment is going to have a significant impact on the development of AI in this area. Only time will tell how much of an impact this will be.
• You can call Maria Peyman on 01223 326596 or email her at maria-peyman [at] birketts.co.uk
