The US Supreme Court’s decision to reject the bid for AI to be inventors on patents could have a ripple effect on UK productivity if the government fails to act.
In Washington DC last week, the US Supreme Court threw out computer scientist Stephen Thaler’s latest bid for AI to be recognised as the legal creator of two inventions. You might wonder why this matters, especially here in the UK.
The argument goes that to be competitive the UK needs to embrace AI and adjust patent law accordingly to serve its original purpose – which is to encourage and incentivise innovation and investment in innovation.
If you only reward companies that do innovation the old way, toiling away with reams of paper, prototypes and expensive tests and trials – and tell others they are cheating by using AI and don’t merit or need similar protection – you risk our ability to make breakthroughs that change lives.
Patents Act is outdated for AI
Today, we operate under the UK Patents Act 1977. These laws were written many years earlier, in a time that could hardly imagine today’s world. It is easy to contend that the Patents Act is no longer fit for purpose, but there seems to be no urgency to change it.
Harvard Law professor Lawrence Lessig and other academics who have supported Thaler are right. It really will jeopardise billions of dollars, pounds or any other currency invested in innovation – as well as potentially lives. When we hear the government touting its vision of the UK becoming a science and technology superpower, there are fundamental building blocks required to reach that goal. Legislation is one of them.
A patented invention is merely the start. The inventions then have to be manufactured and brought to market. A patent itself doesn’t change the economy that has evolved around the need for people to make and sell things to customers.
There’s a theory that every invention is inevitable – it’s just a question of how fast we innovate. AI makes it faster. Patentable inventions must be new, non-obvious and enabled, in that we must be able to build them. So why do we care how it was derived?
Dragging feet will harm UK economy
Prime Minister Rishi Sunak launched a £100m generative AI taskforce last week and promised that investing in these technologies will “build a better future through advances in healthcare and security”. But in the AI regulatory white paper published in March, intellectual property (IP) was given decidedly scant regard.
I predict the problem will arrive in the near future in the field of big pharma and the life sciences sector, on which a good deal of economic growth is predicated. Artificial intelligence or machine learning models are increasingly being used to perform drug discovery. Less than a fortnight ago, Moderna and IBM announced a collaboration to use generative AI and quantum computing to advance mRNA technology – which was behind the pharma company’s Covid vaccine.
The pharmaceutical industry only exists if there is strong patent protection to reward the risk of billions of dollars – typically it takes around $1bn and over a decade to bring a new drug to market. The risk isn’t in the money spent inventing and filing a patent application but instead in the risks of getting regulatory approval over many years of clinical trials, most of which fail. AI can accelerate and de-risk this.
Patents allow companies to recover R&D costs, ring-fencing their ability to generate money from an invention for a limited period of time. Imagine the impact if a patent for a cure for breast cancer failed because it turned out there were no genuine human inventors, which would probably be revealed in litigation first rather than on filing the patent. The repercussive effects on big pharma would surely force legislators to grasp the issue.
And it’s not only big pharma. The telecommunications industry is another where innovation has relied heavily on IP to protect the economic benefit of the companies that have revolutionised mobile communications.
Premature or priority?
At this very moment, there aren’t a huge amount of AI-generated inventions in pharma, telco or elsewhere. The lack of haste suggests those in power think the drive to adapt the law is premature. This is right at one level, but why not do it given the inevitability? It could be five to ten years away, but it could take that long to change the law. Why not do it now? Think again about that cancer-cure scenario.
The data shows that AI applications are growing exponentially in several ways. First, there is AI-implemented innovation, such as the driverless car where the invention is the software. Second, there is AI-assisted innovation, where an AI platform helps come up with a new invention, such as in the field of drug repurposing.
The third is AI-generated innovation, where somebody claims “the machine did it on its own”, as Thaler did. His case wasn’t about the patentability of AI but the patentability of AI-generated inventions.
Rewrite IP laws for AI inventions
The decision in the US was not unexpected – it was effectively a ‘philosophical’ test case or experiment. Judges can interpret the statute but they can’t make it, so without wiggle room to interpret the law, it was unlikely to pass. But it is significant.
US law is all about rewarding inventors and the inventorship is accompanied by a solemn declaration and oath naming all the inventors. Lying potentially renders that US patent as unenforceable. Just as in the US, the law needs to be rewritten here in the UK if we are to entertain the idea of AI as an inventor.
Sunak, along with Tech Secretary Michelle Donelan, will need to take a greater lead on advocating for the sort of changes that would be required to future-proof the innovation economy both here and on a global level.
When you see that a far smaller economy such as South Africa has simply waived this through as a non-issue while we wring our hands, it suggests it’s a less complex and weighty matter than some believe. What happens next could critically alter the UK’s standing in science and technology.
Peter Finnie is a patent attorney and partner at IP law firm Potter Clarkson.