LSPN Europe: AIs should be named inventors, argues Dabus professor
The team leader behind the world’s first artificial intelligence (AI)-generated patent application has said AI inventors will force major changes to the patent system.
Speaking at Life Sciences Patent Network Europe in London today, November 21, Ryan Abbott, professor of law and health sciences, University of Surrey, said the ever-advancing tech behind AI was forcing us to confront issues of inventorship.
In August, the researchers filed an application for the invention, called Dabus, which is said to have designed a type of plastic food container and a flashing beacon light.
“In ten or 20 years, it may be a routine thing that you ask a computer to solve a problem in lots of areas to research. Where using an inventive AI is the standard way to research, I think that means the ‘skilled person’ is really an inventive machine,” Abbott said.
This posed significant challenges for traditional patent law as we know it, Abbott said.
“This really changes the standard because what does an inventive machine find obvious? At that point, we really have to change the test and look more at economic factors.”
Abbott argued that patenting AI-generated works and inventions was the best way of preserving the “primary economic rationale” behind the patent system.
“Namely, it incentivises innovation,” he said. He also argued that recognising AI as an inventor was necessary to protect human inventors as well.
“When an AI invents we should list the AI as the inventor, to protect human, moral rights. Because, if I can ask Siri to invent a thousand things for me and list myself as an inventor, it devalues what it means to be a human inventor.
“It equates the work of someone asking a machine to solve a problem with someone legitimately inventing.”
“Of course, the AI would not own the patent. I’m not aware of anyone actually saying an AI should, though I am aware of a bunch of people strongly reacting to the idea,” said Abbot.
“AI lacks both legal and moral rights, so it couldn’t own a patent. And moreover, there’s just no good reason at all for an AI to own a patent. It has a lack of motivation and an inability to enforce something like that.”
Humans left behind
Because AI’s intelligence will improve exponentially, humans will “have a really hard time” keeping up, Abbott said.
“It’s not that hard to imagine that an AI could sequence every possible combination of amino acids that could make an antibody and publish every possible antibody structure.
“You couldn’t get a patent on those without knowing their utility, but if it’s an anticipatory disclosure it could wipe out compound patents on every future antibody.”
For the most sophisticated computers of the future, “everything will be obvious,” he added, making it harder and harder for things to be ‘non-obvious’.
New laws needed
It is imperative to update laws because AI is going to change fundamental issues of patent subsistence, inventorship and ownership, he argued.
“If what we want from patent law is to incentivise innovation, it doesn’t matter if it’s a robot, a room full of monkeys or Einstein that’s coming up with this stuff; all that we want are laws and policies that result in more innovation. And we will have that whether it’s AI or people that are doing it.”
Did you enjoy reading this story? Sign up to our free daily newsletters and get stories sent like this straight to your inbox.