The director general of WIPO says that questions of
liability for artificial intelligence should be linked to IP ownership as
technological developments begin to change established laws.
Francis Gurry suggests in an interview with Managing IP that
the onus could fall on creators and IP owners in situations where the traditional
liable party is no longer part of the equation.
Pointing to driverless cars as a hypothetical example, Gurry
says international conventions would usually point to liability in the event of
an accident resting with the driver.
“Remove the driver, and who is responsible?” Gurry asks.
“It’s a complex question and depends on what went wrong. The person who created the AI and put it into operation has the responsibility, I think.”
Gurry spoke to Managing IP yesterday during the AI: Decoding
IP conference in London.
He also says he does not see why traditional IP rights
should be attributed to non-human entities, adding that liability and property
rights should be connected in a future world.
“It is right that there should always be a human at the end
of it,” he says. “Look at automated cars or drones, if you have a machine or
algorithm organised in certain way and you attribute the rights resulting from
that to a machine, what do you do if the drone or vehicle crashes into a primary
school?
“Liability is connected with property rights.”
He adds that he does not see much appetite to attribute
rights to machines from an innovation standpoint either: “Why do we attribute
rights? One is the moral reason of giving just reward for someone who has
created something. But the main reasons are economic, because we want to
encourage others to develop innovation.”
The full interview, in which Gurry talks more about how to
determine property rights in the AI world, how technology has improved patent
and trademark filing systems, and how the use of data can be linked to IP, will
be published on Managing IP next week.