AI systems cannot be listed as inventors in patent submissions, the US Patent and Trademark Office ruled this week.
Last year, the USPTO received an odd patent application. The sole inventor on the submission, describing a method for “attracting enhanced attention,” had the first name “DABUS” and the family name “Invention generated by artificial intelligence.”
The assignee – the person who has the property right to the patent – was listed as Stephen Thaler, founder of a Missouri outfit called Imagination Engines. The paperwork was filed in the UK, Europe, and Israel as well as the United States.
The patent application outlined [PDF] a gadget that emits light with a fractal pattern, and could be used as, say, an emergency beacon.
The idea was, apparently, invented by a set of neural networks dubbed DABUS, which was created by Thaler. He said it was only right to give credit where it was due, namely his software.
USPTO officials, however, declared the application invalid because it did not “identify each inventor by his or her legal name.” Thaler was subsequently fined $80 for an incomplete submission.
In response, he filed a petition requesting a probe into how his submission was handled, as he believed the wrist-slap was unwarranted. USPTO filed a second notice reminding Thaler his application was still missing vital information. Both parties continued to duke it out until the USPTO decided to give Thaler just two more months to refile his case, this time with a human inventor.
In a detailed decision, deputy commissioner for patent examination policy Robert Bahr said: "Interpreting ‘inventor’ broadly to encompass machines would contradict the plain reading of the patent statuses that refer to persons and individuals.”
USPTO officials also reiterated: "Under current law, only natural persons may be named as an inventor in a patent application."
AI software should be able to register its own patents, law prof arguesREAD MORE
Thaler, however, believes being able to credit machines as inventors would encourage more innovation in artificial intelligence, and prevent people from stealing ideas generated by computers.
“As AI begins to crank out progressively more inventions, humans will watch and be tempted to file patents that they really didn’t conceive,” Thaler told The Register.
“Of course there would be consequences for such dishonesty, especially if appropriate declarations have been signed. [It] will likely lead to a lull in technological advancement, with intellectual property output becoming only a small fraction of what it could be. The collateral damage could be a storm of intellectual property streaming from the countries we are competing with scientifically and technologically.”
The USPTO was not swayed by his arguments. It referred to previous rejected applications where the inventors were states or corporations.
The ruling follows similar decisions by the European Patent Office earlier this year. Thaler had filed two Euro patent applications for a food container and the aforementioned light beacon, which were dismissed because the listed inventor – DABUS – did not have a full name and address nor the ability, as an AI, to own property or transfer ownership to Thaler.
What happens when a highly advanced extraterrestrial civilization visits Earth?
“Humans are denying those rights in the first place, being stuck in an age-old paradigm rut in which only wet computers – ie: brains – count. What happens when a highly advanced extraterrestrial civilization visits Earth?" Thaler told El Reg.
"Does ET deserve the equivalent of human rights? Does he dare file for patent or copyright protection? Then what happens when science achieves the download of consciousness into machines, or silicon prostheses are introduced into the protoplasmic brain? Would the system deny these people human rights? …Those are mighty big questions.”
Thaler does not believe any old algorithms should be accepted as inventors: the software has to be advanced like DABUS, which stands for Device for the Autonomous Bootstrapping of Unified Sentience. He claimed the software could “not only formulate new ideas, but also appreciates the repercussions of what it conceives, as well as developing subjective feelings for its breakthroughs.”
Who'd have thought artificial sentience would be developed not in the halls of Stanford, MIT, Berkeley, Cambridge, Google or Microsoft, but in Saint Charles, Missouri.
In any case, USPTO has given Thaler a couple of months to resubmit a valid application listing a real-life person for the invention.
"With respect to the present applications we will not be doing that because there is no person, at least according to traditional criteria, who we believe qualifies as an inventor," Ryan Abbott, professor of law and health sciences at the University of Surrey, England, who is representing Thaler, told El Reg. "We plan to appeal the decisions." ®
Sponsored: Webcast: Ransomware has gone nuclear