Dawn Ellmore on the world’s first patent applications for AI-designed inventions


Dawn Ellmore heads up Dawn Ellmore Employment, a specialist patent, trade mark and legal recruitment consultancy. With more than three decades’ experience working within Intellectual Property (IP), Dawn has seen many changes. But perhaps none as potentially game changing as the first patent application filings for inventions created solely by artificial intelligence (AI)…

A team of academics from the University of Surrey has filed the first ever patent applications for AI-created inventions. That means no human inventor contributed to the development of the invention, and the patent applications are under the name of the AI inventor – DABUS. 

DABUS (Device for Autonomous Bootstrapping of Unified Sentience) is the creation of Stephen Thaler, a pioneering AI researcher based in Missouri. According to the Financial Times, Mr Thaler taught DABUS to produce ever more complex items using words and images.  

What are the first patent applications for AI-designed inventions? 

The artificial intelligence inventor uses a complex maze of interconnected neural networks to come up with its own ‘ideas’ to solve problems. It does this by varying the connections between the neural networks. A secondary layer of equally complex neural networks works in the background to predict the critical consequences of the invention. 

So, what has DABUS come up with? The output generated by the AI forms the basis for two patent applications. The first is for a shape-changing plastic food container based on fractal geometry. The other relates to a flashing light device that can be used to attract attention during search and rescue incidents. 

The two AI-designed patent applications were filed by a team headed by Professor Ryan Abbot, Professor of Law and Health Sciences at the University of Surrey. In an interview with Deezen, Professor Abbot says he is seeking “clarity surrounding laws and rules of AI generated inventions.” 

Lack of clarity regarding patent accreditation 

AI has fundamentally shifted from being a sophisticated tool used by humans to predict, improve and develop, andit is now automating innovation, which has the potential to profoundly alter human progress. 

Both the UKIPO and the European Patent Office (EPO) accept that the inventions made by DABUS are eligible to receive a patent. In simple terms, this means that the light device and the container are considered to be industrially applicable and brand-new inventions. However, the fact that the inventions are not the product of human development opens up a whole new world for patents. 

For example, it remains unclear who (or what) will be credited as the owner or holder of the patent, should it be granted. There are no laws in any country in the world to specify how cases like this should be dealt with. And while AI has been on the global radar for decades as the future of creativity, there is no precedence for an AI machine to be granted a patent or to be credited as an inventor. 

Can AI be granted patent ownership? 

Professor Abbot says that the AI does fulfil the criteria that forms the basics of inventorship in these two patents: “There would be no question the AI was the only inventor if it was a natural person.”

One approach could be for the AI to be credited on the patent as the inventor, while the owner of the AI is given ownership of the patents themselves. If this goes ahead, the patent system would continue to support invention and innovation. It would allow the development of more inventive and more sophisticated AI, and bypass obstacles regarding ownership. 

However, according to a study presented to the EPO regarding inventorship and AI inventions, experts think this unlikely. Noam Shemtov, Deputy Head of Commercial Law Studies at Queen Mary University of London, doesn’t expect the legal position regarding AI inventors to change in the short-term, saying: “At present there are no convincing reasons to consider a change in this respect.” 

Global IP law on AI must change

There is no doubt that AI could hold the key to some of the biggest, existential crises facing humanity. Whether AI systems eventually discover a cure for cancer or develop workable methods to reverse climate change, the ramifications of this technology are far-reaching. 

But there is also no doubt that IP laws must catch up with the rapidly changing pace of technological advancement. As Professor Abbot says: “If outdated IP laws around the world don’t respond quickly to the rise of the inventive machine, the lack of incentive for AI developers could stand in the way of spectacular human development.”

The two patents in question will be examined following their formal publication in April and May next year (2020). Applications for these two machine-created inventions are pending at the UKIPO, as well as at the EPO and the United States Patent and Trade Mark Office (USPTO).  

Currently, the UK Patents Act of 1977 and the European Patent Convention both say that inventors can only be ‘natural persons’. In the US, the law says inventions must be made by an ‘individual’. 

By making these patent applications, Professor Abbot hopes to bring the lack of clarity regarding patent accreditation to light. He says that innovation must be rewarded properly regardless of the source of the invention. 

About Dawn Ellmore Employment 

Dawn Ellmore Employment is a market leader in intellectual property and legal recruitment.