Patent lawyers warn of AI’s ‘double-edged sword’

Patent lawyers are warning that, while the adoption of artificial intelligence tools may well deliver dramatic savings in time and efficiency, it could also weaken the protection of new inventions.

AI is set to radically change the traditional role of specialist patent lawyers, which has been central to bringing innovative devices and products to the global marketplace.

Already, machine learning is cutting the workload of patent attorneys, who mainly use the technology to automate laborious manual tasks. These include patent searching, gathering technical information, and drafting and filing applications.

In a 2023 survey of 575 intellectual property and research and development professionals, conducted by analytics consultancy Clarivate among global firms, many IP practitioners said they had seen significant adoption of AI in their field. At the same time, though, 74 per cent of those surveyed expressed reservations, with accuracy being their top concern.

Some lawyers warn that AI tools could now be used to build cases to invalidate patent applications — with potentially damaging effects for the R&D sector.

“AI is finding documents that people would not have found because it has more processing power,” explains Anthony Albutt, partner at UK-based D Young and Co who leads the firm’s mechanical patent practice in London. “That is important because, if I want to invalidate a patent, these documents can be used to present a case to a judge as to why it should not have been granted.”

While it might appear to be a good thing that information is easier to find, this can get in the way of progress. “If many more patents are revoked because of the sudden availability of a vast list of documents, it changes the economics of investing in technology and using patents to protect that investment,” argues Albutt.

As such, AI may be a “double-edged sword”, he suggests. “It’s going to make lots of patents very vulnerable”, which could have “a hugely detrimental effect” on innovation.

Vasheharan Kanesarajah, Clarivate’s head of strategic development for intellectual property, agrees that a key challenge for practitioners is confidentiality and the disclosure of inventive concepts via AI search tools.

For example, if an IP lawyer or inventor, while conducting public patent database research, types an inquiry into an AI tool such as ChatGPT, that might make an invention commercially vulnerable. “[When] you start to disclose a lot of information into an open access AI tool,” Kanesarajah explains, “that changes everything and creates a lot of risks for corporations.”

In response to these risks, lawyers will need to adapt, skill-up, and get better at using AI programmes to maintain a competitive advantage — especially for patent research, where AI tools are proving most useful to the profession.

“It’s not a one size fits all,” says Kanesarajah. Learning to use AI tools to find the right data requires instruction from a legal specialist in that particular field, whether it be rocket science or telecoms, he points out. “The attorney will need to know how to train, query or develop prompts for [AI programmes]. That is a skillset that they are getting better at.”

As patent lawyers adapt to a more complex world, therefore, doomsday scenarios in which they are replaced by AI seem unlikely. Rather, it could be argued that their roles are as likely to expand.

As a growing number of creations — from artworks to vaccines — are produced either by, or with the help of, machine learning, lawyers will have to prepare for a shake up. With the speed of innovation increasing, there is set to be more demand for the drafting, filing and negotiating of patents — putting pressure on lawyers to keep up, says Kanesarajah.

Currently, the three leading global sectors for AI-related patents are telecommunications, transport, and life and medical science, according to research by the World Intellectual Property Organization, a global forum for intellectual property services. Patent lawyers in these fields, in particular, may need to rise to the occasion.

Practitioners are also closely tracking debates on “inventorship”. Under US and UK law, inventors must be human, rather than machines, although this is being challenged in the courts. And these questions over whether present-day law is fit for purpose are creating new considerations for patent lawyers in cases where AI tools are used.

In one recent high-profile case, the UK Supreme Court ruled that a trained neural network, known as Dabus — a type of machine learning model designed to mimic the human brain — could not be listed as the inventor on two patent applications filed at the UK Intellectual Property Office. But this judgment raised speculation over how countries might look to change their laws as AI advances.

Avi Freeman, a partner and patent attorney at BeckGreener — which specialises in physics and electronics patents — argues that, despite the transformations sweeping through the profession, the role of specialist lawyers will not disappear. It will be more a case of the next generation acquiring different skills from the last, he predicts.

“It’s not that AI will replace lawyers,” he says, “but rather lawyers will be replaced by lawyers who know how to use AI.”

Letter in response to this article:

Patent lawyers are guilty of disingenuous AI claims / From Andrew Fentem, Oxford, Oxfordshire, UK