In recent years, the rapid advancement of artificial intelligence (AI) has led to the emergence of complex questions surrounding intellectual property (IP) rights.
As AI technologies continue to evolve, the need for a comprehensive legal system to address these issues becomes increasingly apparent. The Intellectual Property Office (UKIPO) carried out a consultation in 2021 which specifically considered how computer-generated content is dealt with in this jurisdiction, with the aim of recommending changes to our current legislative framework, if necessary.
However, the government has subsequently decided to shelve the recommendations that came out of the consultation and take a ‘wait and see’ approach. Whether this is sensible, or just creates more uncertainty, remains to be seen. Most commentators seem to think that the recent Safety Summit at Bletchley Park has done little to clarify the UK’s approach.
What we do know is that there are (at least) two critical questions from a legal perspective which are relevant to the development of both AI and IP law which need to be dealt with (and clarified):
-
Who owns the content created by AI?
-
What rules govern the use of content by AI, particularly when it gathers information from various sources that have the benefit of copyright protection?
While we await the development of legislation, codes or case law to clarify these matters, we will analyse potential outcomes and examine the current situation in UK.
Ownership of AI-Generated Content
Unlike most jurisdictions, English law does not require there to be a human ‘author’ of a copyright ‘work’. Section 9(3) of the Copyright Designs and Patents Act 1988 (CDPA) which deals with the ownership of computer-generated works, states that the author of a computer-generated work is deemed to be “the person by whom the arrangements necessary for the creation of the work are undertaken.”
This should mean that the individual who initiates the process that creates the AI-generated output is considered the author of any resulting copyright work, following this rule. If that person is an employee, then following the usual rule, the employer will own the output that is created. However, this has not been tested in the English courts at the moment, so we do not know how section 9(3) of the CDPA will be applied to AI models. One obstacle could be the requirement under English law that a work that is created needs to be original to benefit from copyright protection. The concept of originality has evolved over the years and case law dictates that it requires the author’s own intellectual creation. That is difficult when the creator of that work is not a human and the ‘author’, (in accordance with section 9(3)) is simply the person that is prompting the AI tool.
Rules Governing AI and Content Usage
The second question is around the rules governing AI’s use of content to gather information or create textual and visual outputs. This issue is currently under scrutiny in the case of Getty Images v Stability AI. Getty Images alleges that Stability AI has unlawfully copied and processed copyrighted images and metadata without the necessary licences, benefiting its commercial interests while harming content creators.
The outcome of this case could set a precedent and establish the legal boundaries for AI’s content usage. If Getty Images prevails, it may discourage AI innovators from scraping the internet for content without proper authorisation. This could be contrary to the government’s vision to grow the artificial intelligence industry in the UK. It originally intended to do this by enabling text and data mining by AI through appropriate copyright laws which would allow better access to AI systems to enable machine learning. However, that plan has now been abandoned, which is a disappointment for advocates who argue that easing copyright restrictions on AI could promote wider AI adoption.
Opponents of this view allege that copyright and related rights do not obstruct AI development. They argue that licensing models can adapt to evolving technology, providing AI developers with opportunities to license copyrighted works. Instead of limiting copyright protection, they think that the focus should be on ensuring fair compensation for rights holders when AI utilises their works.
Conclusion
In achieving its goal to place the UK as a global hub for AI and data-driven innovation, as we have said, the government at first seemed inclined to support innovation and adopt a data mining exception (for commercial purposes) that would at least help with the issue of unlawful copying when the AI tool processes copyright-protected works and generates an infringing work as a result.
However, as the recommendations flowing from the UKIPO’s consultation have been paused, partly as a result of the concern regarding the effect that this would have had on the creative industries, the UK is taking a different approach from that of the EU and USA which intend to legislate to try and deal with these complex issues.
The recent summit at Bletchley Park has not done a huge amount to clarify the UK’s approach to AI (certainly from a regulatory standpoint). Obviously, this was a ‘safety summit’ with the clear intention of addressing safety as a specific issue. However, one would think that safety and regulation go hand in hand.
As AI continues its evolution, the legal landscape must adapt to ensure clarity, fairness, safety and innovation. The coming years will undoubtedly see many developments in AI and IP law, shaping the future of AI technology in the UK.
If you are concerned about AI and copyright issues, and would like to speak to someone with expertise in the sector, why not contact our Intellectual Property team? Approachable and highly experienced in all areas of IP law, they are ready to help you with whatever matters you are facing.