A recent TechRadar article explores the evolution of AI technologies that could conceivably outperform humans in creative disciplines previously perceived as uniquely human. The proliferation of AI technology across a wide range of sectors and industries evidently poses a number of challenges for intellectual property rights historically designed to protect human creativity and innovation. Whilst AI is unlikely to completely replace human creativity in the near future, there are pressing questions about how intellectual property law interacts with AI technologies and AI-generated output.
One of the ongoing topics of debate is whether AI is capable of creating and owning protectable subject matter. In the copyright sphere, the law already permits computer-generated works to be protected by copyright, however the author (and owner) of the protected work is the person “by whom the arrangements necessary for the creation of the work are undertaken”. Whilst copyright recognises protection of AI-generated works there seems a reluctance to allow non-human ownership of those works. Arguably by designating a human as the author of AI-generated works, the law fails to recognise the reality of truly independent AI-generated works. Further, the current legal principle effectively divorces authorship and creativity which sits uneasily with the general understanding of originality where creativity and authorship come hand-in-hand.
Another current topic of discussion is the extent to which it should be permissible for AI to scan whole databases of protected content owned by others in order to learn and develop new products. To a certain extent the process of scanning large datasets is fundamental to the way AI functions and its capacity for analysing such large datasets is one of the core benefits of AI technology. But that very process carries the risk of infringing database rights and potentially underlying copyright subsisting in third party works. UK law currently has a narrow exception to copyright infringement for text and data mining, but it is limited to non-commercial purposes and is confined to potentially infringing copies within an AI system, rather than aspects of copies embodied in the output from it. A key area of debate is whether the copyright framework should make it easier for AI technologies to use protected subject matter, and this is a fundamental aspect of a UKIPO call for views on the implications of AI technology. The consultation has closed but the UKIPO is yet to report on the responses so it will be intriguing to see the findings on this issue and many other AI related issues.
Thinking about the long term effects AI might have on intellectual property law, it is conceivable that AI products (smart home products like Amazon’s Alexa and Google Home spring to mind) could fundamentally change consumer purchasing habits. These forms of AI could effectively replace humans in the purchasing process, which could in turn have implications for fundamental concepts of trade mark law. For example, if purchasing decisions are left to AI products what implications will that have on the concept of the average “consumer”, a concept which sits at the heart of trade mark law? Similarly, if humans are removed from the purchasing process, will post-sale confusion become the new norm of infringement? It may be that shifts in human behaviour do not impact certain market sectors, for example high value purchases, but commodity products seem far more vulnerable to a shift in purchasing habits driven by AI technology. For the time being these issues are largely for academic debate in IP blogs, but clearly when the time comes the law will need to evolve to account for the implications of developments in AI technology.
There are obvious commercial benefits to be gained by harnessing the power of AI. Yet there are numerous uncertainties around the application of intellectual property law to AI technologies and AI-generated outputs. Whilst businesses may be able to benefit from the application and use of AI technologies those utilising the technology should be mindful of the question marks around the subsistence and ownership of IP rights in the output from AI technologies. Likewise, businesses should be conscious of the potential infringement risks inherent in the way AI technologies process data and content and also risks with AI-generated outputs which could be derived from other protected subject matters.