AI Can’t be Named as an Inventor—Yet

  • Added
  • Author:
Artificial intelligence is offering many benefits to the intellectual property (IP) industry, but it can’t legally take credit for inventing anything new. A patent application listing an artificial intelligence (AI) entity as the inventor was denied by three major patent offices. Submitted by Dr. Stephen Thaler with the inventor listed as DABUS, the USPTO, the EPO, and the UKIPO, all agreed that current law can’t allow for anything other than “natural persons”. Dr. Thaler purposely submitted the application knowing that it would stretch existing statutes and expand discussion about AI creativity.
While the reasoning was similar for all three of the patent organizations, the comments released with the judgements differed in a few distinct ways:
  • The UKIPO rejected the application because the naming of a machine as inventor (DABUS) does not meet the requirements of the Patent Act, stating that a person must be identified. The agency noted that AI will play an increasing role in the inventive process and the question of inventorship is worth debating for the future.
  • The EPO stated that new legislation is required to give AI credit as an inventor, in which a “legal personality” would be established for the AI entity and assigned to a person.
  • The USPTO decision didn’t mention any possibilities for the future of AI inventorship, but opened up discussion on the issue through a Federal Registrar Notice.
DABUS is short for “Device for the Autonomous Bootstrapping of Unified Sentience” and is a patented invention itself. Its primary claim is to be a self-learning, artificial neural network-based discovery device. The entity is pre-dated by another patent granted in 1994 called the “Creativity Machine”, which is designed to create new inventions by ingesting large amounts of related material, similar to the way traditional machine learning works. DABUS goes a step further, claiming to come up with new ideas without any specific input.
Yet the use of AI in making discoveries is nothing new. Creative output from computing started in the 1980s, according to a paper published by the Boston Law Review written by Ryan Abbott. Many patents are already granted to inventions made by computers but still use a person’s name as the inventor. Abbott thinks that by allowing computing technology to take credit for new technologies in patents and other IP, more effort will go into creating inventive computing and those new entities will be able to solve some of the most challenging problems of science.
Whether or not creative computing gains the right to be listed as a patent applicant, the growth of AI technologies continues to expand. Our fastest growing technologies report issued in January 2020 shows that patent applications for AI have grown by a compound growth rate of 42% from 2014–2019. AI will clearly play an increasing role in future inventions, whether it can legally be listed as an inventor on a patent application or not.