‘A country will soon accept AI inventor patents’ says DABUS patent attorney

Panellists from Qualcomm and Clarivate discuss AI trends in key Future of IP webinar | The West would be “scrambling to change laws” should China push ahead with AI inventorship | Scepticism could lead to “insurmountable barriers” to protection, says Qualcomm chief IP counsel.

The patent attorney who wrote the applications for DABUS has predicted that “a country of significance will soon accept the grant patents for AI-generated inventions”.

Robert Jehan, partner at Williams Powell, told a WIPR webinar that the “West is going to be playing catch up” should a country like China make good on its intentions.

Speaking on AI and IP: What happens next?, part of this month’s Future of IP focus, Jehan said a senior China judge has already indicated that there's no barrier to granting patents for AI-generated inventions in the country.

“We keep on saying we want to be leaders in AI, you don't lead from the back,” said Jehan, who authored the UK and EPO applications for the DABUS, the artificial inventor project of legal test cases seeking IP rights for AI-generated output.

He warned that should China press on, Western countries would be “scrambling to change our laws, to change the USPTO guidance that has recently come out, to reverse the balance that they're now trying to impose.”

'Very big unanswered questions' 

Jehan joined a webinar panel with Arun Hill, senior consultant for the Clarivate Center for IP and Innovation Research, and Bob Giles, senior vice president and chief IP counsel at Qualcomm.

In Hill’s view, the impact of AI is overstated in the short term, but in the long term, it's underestimated and understated.

He called for deeper engagement with fundamental policy questions regarding the patent system.

“There are still some very big unanswered questions so it will be interesting to see the kind of knock-on consequences of conversations from AI regulation, such as the EU AI Act, and how they will filter down to the level of a patent law and even public practice,” said Hill.

Giles, whose team helps protect leading edge AI-related hardware and software, emphasised the need for more attention to be given to patents.

“I do think that there's a lot of attention being given to the copyright side, and especially on fair use and other issues that are coming to light on training data, but I think that very close attention needs to be paid on the patent side, as well to prevent the disincentivising of companies to do the true R&D in these areas,” said Giles.

He urged policymakers to engage in serious discussions about patenting in AI to prevent unfavourable outcomes that could be difficult to reverse.

AI inventorship and the UK Supreme Court decision

Jehan wrote the UK and European Patent Office applications for the DABUS device, an invention made by an AI system that has tested the rules of IP offices the world over.

Jehan noted that the project influenced the US Patent and Trademark Office’s (USPTO) guidance, which acknowledged AI's role in invention.

Currently, Saudi Arabia and South Africa (although the country does not have substantive patent examination) have accepted the naming of DABUS as the inventor on those patent applications.

There are appeals still pending in the US, Japan, China, Korea, India, Israel and in other jurisdictions, however, a final decision was reached in Australia.

Jehan called the UK Supreme Court’s ruling, focusing on whether a person could be named as the inventor when the invention was made by an AI system, “narrow”.

"The Supreme Court said had they been given a different question, the outcome may very well have been different,” noted Jehan, who pointed to a surprising aspect of the ruling.

The Supreme Court implied that even if Thaler had named himself as the inventor, he wouldn't have had rights to the patent, raising concerns for industries reliant on AI-generated technologies, said the patent attorney.

Jehan described this issue as necessitating an “urgent change” in UK law.

Jehan: Patents ‘do not reward the inventor’

Jehan delved into the issue of named patent applicants needing to be human, a common reason given by IP offices for rejecting DABUS applications.

He noted that the UK IPO said in a submission to the UK Supreme Court that it doesn’t matter who you name on the application because it's actually the nature of the invention, which is important.

“That's really where we ought to be going” said Jehan.

Jehan said language of ‘rewarding the inventor’ used in DABUS rejections, and the USPTO’s AI guidance, missed the point.

“Patents do not reward the inventor, patents reward the employer—the owner of the invention. [The owner is] the entity that has to make the huge investment most of the time, in marketing the invention, commercialising the invention and all the rest. That's why the patent system existed,” said Jehan.

The patent system also exists to make the invention public, he said.

“[A patent is] actually a reward for publishing an invention, and for companies like Qualcomm to make a huge investment in developing new technologies, and to give them some form of protected period where they can recoup that investment,” added Jehan.

“We're at a situation now where we need to think beyond how the invention was created to whether or not it is an invention that is worthy of patent protection.”

On this topic, Giles expressed concern about the emergence of “unreasonable barriers or obstacles to patenting” in light of evolving AI technologies, and worried that scepticism towards AI-generated inventions could lead to “insurmountable barriers” in obtaining patents.

While some guidance seemed promising, the Qualcomm legal exec emphasised the uncertainty surrounding its interpretation and application by patent offices and courts, and hoped for clearer guidance and solutions that wouldn't undermine the patent system's integrity.

Challenges of protecting inventions with an AI element

Giles highlighted several key challenges, including the need to keep pace with the tech and any legal developments.

"One of the biggest challenges is keeping up with everything that's going on right now," he said.

Additionally, he emphasised the importance of thoroughly explaining the technical aspects and benefits of inventions in patent specifications to address algorithmic concerns and ensure clarity for examiners.

Giles explained: “We need to make sure we're explaining the specification, the purpose of the invention, and the technical effect very well so that the examiner can understand that it's much more than just an algorithmic application.

"You only get one chance to submit the specification right to the patent office," he added, so “explaining the technical nuance and the technical benefits of the invention as clearly as possible up front" is important.

Giles also discussed divided infringement, where multiple actors may be involved in implementing the invention, and the necessity to navigate AI assistant inventions and their associated requirements from patent offices and courts.

Despite AI's relatively limited current use in the inventive process, there is increasing attention from patent offices, which Giles views as beneficial for patent applicants seeking guidance to assert their rights effectively.

"It's good to see that the patent offices are taking this action seriously and trying to provide some guidance,” said Giles.

"As much guidance as we can get up front, the better it will be when we try to assert our rights."

Use of AI by legal professionals

Hill discussed the legal sector's adoption of AI tools, highlighting notable results from a 2023 Clarivate report.

He noted that while scepticism of tools, such as generativeAI co-pilots, exists, it's rooted in a precautionary approach and evidence-based reasoning typical of the legal industry.

While he does not use AI yet in his work, Giles is “assessing the tools carefully” for potential future use.

He believes they are “nowhere near” the stage of being able to generate patent applications, but he noted that certain tasks, particularly time-consuming and mundane ones, could benefit from AI assistance.

“Even with some of the more mundane, laborious tasks, AI tools are getting close, but they're not quite there yet,” said Giles.

“It takes quite a bit of finesse to it, understanding and modifying the results, so it definitely needs to be curated by attorneys and by people that have worked in the industry for quite some time,” argued Giles.

Jehan echoed Giles on this point, but also acknowledged the potential usefulness of these tools in maintaining consistency with terminology and structure.

“One of the most important things that pattern drafting is the intimate discussion between the patent attorney and the inventors to try to tease out from the inventors what the real core invention is.

“Sometimes you're acting as a trigger to get the inventors to think further and beyond where they have done, because what you're trying to achieve is something that hopefully would last 20 years, " Jehan added.

An AI is nowhere near able to do that, he concluded.