USPTO Rules That AI Cannot Invent Patents

The U.S. Patent and Trademark Office (USPTO) ruled last week that artificial intelligence systems cannot be credited as an inventor in a patent. This decision was in response to two patents created by an AI system called DABUS.

The USPTO cited that U.S. patent law refers to inventors with humanlike pronouns, including “himself” and “herself.” The group that filed the patents argued that because the law references inventors as “individuals,” it could be applied to machines. The USPTO concluded that this understanding was too broad, saying, “Under current law, only natural persons may be named as an inventor in a patent application.” The UK’s Intellectual Property Office (IPO) and the European Patent Office (EPO) have also ruled that DABUS cannot legally be considered an inventor.

The Artificial Intelligence Project seeks intellectual property rights for the autonomous creations of artificial intelligence. Dr. Stephen Thaler, a member of the group, created DABUS. It is a patented AI system dubbed a “Creativity Machine,” referring to its specific kind of artificial intelligence. Systems like DABUS contain branches of neural networks that are trained with information from different knowledge sources. They also act as self-critics by monitoring their own neural networks for new ideas.

Ryan Abbott, another member of the Artificial Intelligence Team, believes that changing the legal terms involving patents would incentivize innovation in the AI sector. In an interview with Financial Times, Abbot said, “If you make a point of recognizing how valuable a machine has been in the creative process, that machine will inevitably become more valuable.” As of now, artificial intelligence is considered a helpful mechanism in the inventive process, rather than a sole inventor.