Could an AI system own a patent? Changes anticipated to patent ownership
Patent ownership is at a crossroads. It has long been the go-to intellectual property (IP) right for inventors looking to protect and gain an exclusive right over their inventions, with ownership of a patent often seen as a badge of honour and a valuable asset, writes Raina Victor, Solicitor, Commercial & Technology with Birketts LLP.
However, with the increasingly widespread use of artificial intelligence (AI), it is inevitable that AI itself has started to, and will continue to, create and devise new inventions on its own, without any human input. In this case, who gets to be the “inventor” on a patent application – the AI or the human?
Currently, the law states that, when applying for a patent, the inventor must be: (i) named on the application, and (ii) be a natural person, i.e. a human. This was confirmed by the Court of Appeal last year, which held that an AI machine could not be named as an “inventor” of a patent as it is not a “natural person”.
Instead, the court held that the “inventor(s)” on such applications should be the humans who used the AI to devise the inventions. Interestingly, in other countries, such as South Africa and Australia, AI systems have been recognised and accepted as an inventor on patent applications and AI systems are able to own patents.
This case sparked a debate as to whether law in the UK should allow an AI system to be named as the patent inventor for AI-devised creations. Since the case, the UK government has published two key strategies: the National AI Strategy and the UK Innovation Strategy.
The core focus of these strategies is to make the UK a global AI and science superpower over the next decade and to create the right conditions for all businesses to innovate, in order to attract investment for such innovations. With the protections afforded by patents and other IP rights key to implementing the two strategies, the UK Intellectual Property Office (UKIPO) carried out a public consultation on how patent laws should change, if at all, to protect inventions devised by AI.
The consultation’s aim was to find out if the current IP legal regime strikes the appropriate balance between encouraging investments into inventions, and therefore continued economic growth, without undermining the role of IP in promoting human creativity and innovation.
The consultation proposed the following four options for reform in relation to AI-devised inventions:-
- Make no changes to the current patent laws, which state that patents are only available for human inventors. AI-devised inventions would not be patentable
- Expand the definition of “inventor” to include the humans who made the necessary arrangements for the AI system to devise the invention. This includes those responsible for programming, configuring and operating the AI system that devises the invention. Unlike option 1, AI-devised inventions would be patentable, but a human inventor will still need to be named on the patent application
- Recognise the AI system as an inventor in patent applications. Similar to option 2, AI-devised inventions would be patentable, but in this case, it is proposed that either the AI system could be listed as the inventor on the patent application, or, alternatively, the requirement to name an inventor could be removed altogether for AI-devised inventions; or
- Introduce a new, separate protection for AI-devised inventions with fewer exclusive rights and a shorter duration period than currently provided by patents, which runs alongside the existing human-centric patent system.
The primary argument for reform is that a failure to provide IP protections to AI-devised inventions could discourage research using AI and, as a result, investment into AI.
It is clear that AI can be used to expedite the pace of inventions and reduce research and development costs compared to human inventions. Moving away from the status quo and affording protections for AI-devised inventions would therefore be an effective tool in achieving the Government’s ambition of encouraging innovation in AI, leading to investment into the UK technology sector.
However, there is a risk that adopting options 3 or 4 could potentially lead to competition issues. For example, if AI accelerates the pace of inventions, it could lead to a few dominant players using their AI systems to increase their patenting activities and gain protection over a wider range of inventions.
This would increase the barriers to entry for start-ups and SMEs, making it difficult to innovate without infringement or incurring significant cross-licensing fees.
Another risk is that in most other jurisdictions, only the humans involved are recognised as “inventors” in patent applications. A divergence from this could be detrimental to businesses operating internationally, who wish to apply for patent protection across multiple markets.
The concerns around competition and the ability to patent globally could consequently tip the balance away from encouraging human creativity and innovation.
Although the UK IPO’s consultation ended in early January this year, the findings are yet to be released. With a number of test cases being brought around the world to decide whether AI systems are capable on being an “inventor” of an AI-devised invention, courts across the globe are also at a crossroads.
It will be interesting to see which of the four proposed roads the Government will take and what the future holds for AI-devised inventions.