top of page
  • Rahul Vijh

In re DABUS: Can AI be granted patents?

A landmark decision in relation to the curious case of DABUS AI and Stephen Thaler set a major precedent for patent eligibility for Artificial Intelligence (AI) created inventions. On July 29, 2020, Stephen Thaler submitted a patent request with the inventor’s name as DABUS and listed the family name of the inventor as "Invention generated by artificial intelligence." This was done due to the European patent Office’s (EPO) rejection of the patent claim based on the inventor’s name on the application. In November 2019, the EPO rejected the patent application of Stephen Thaler as it did not comply with the rule which necessitates stating the family name and address of the inventor. However, the EPO did not address the issue of whether DABUS, a non-human intelligence system can be awarded the title of an inventor. In January 2020, after Thaler made corrections to the name of the inventor, the EPO refuted the application once again saying even if DABUS has a name and address, the AI system lacked a legal personality and hence will not be awarded a patent in the EPO.

DABUS is an artificial intelligence system that can ‘invent’. DABUS is a Device for the Autonomous Bootstrapping of Unified Sentience. Dr. Stephen Thaler and his team of researchers spent over a decade developing this AI system which invents without needing to be given specific data or instructions of doing so. Thaler claims that DABUS uses general information fed to it to develop anything in its purview. DABUS invented two applications; one was directed at emitting a blinking light in a manner to cause alerts during an emergency and the other referring to an easier food transporting container.

On April 27, 2020, The United States Patent and Trademark Office denied the patent application by Thaler (as the assignee) with the AI system, DABUS, as the sole inventor. The decision took support from the statute of the United States Code, Title 35 "whoever invents or discovers . . . may obtain a patent, therefore”. The emphasis on the word whoever is to signify that the patent award is to the person who interprets the results and not the one conducting data analysis, organization, etc. Another precedent that the USPTO relied upon was Univ. of Utah v. Max-Planck-Gesellschaft Zur Foerderung der Wissenschaften. The case is elementary as it ruled that only natural persons could be inventors under the U.S Patent Law. The use of words such as ‘mental’ and ‘mind’ indicates the need for the creative processes that result in the patentable invention to be performed by a natural person. The case was originally related to the question of inventorship of states and organizations. The USPTO held that the result will be interpreted in the same manner. The problem of ownership was also raised by the USPTO. As DABUS was not eligible to own property, patent eligibility clashes with the right to own property as well. Hence, it was ruled that machines cannot be inventors. In other words, patent inventors must be human.

Artificial Intelligence is a rapidly evolving technology that will define and accelerate the state of the art for decades to come. The compound annual growth (CAGR) of AI is calculated at 33.1% Market rise by 2026. By the end of 2026, the AI industry is calculated to reach USD 202.57 billion. As often the case with wide-ranging disruptive technologies, the laws will have to develop along with it to encompass AI. Even though the USPTO ruled that only humans can be patent inventors, there are loopholes in the procedure. The USPTO has decidedly not made any determination regarding who or what can create an invention. (In the case of DABUS, the lack of human involvement was highlighted.) The UK Intellectual Property Office (UKIPO/IPO) recognizes even within its decision that the current patent system, scenario, and law may not be sufficient enough to cater to the issue of AI. The applicant has the right to make an appeal in the higher courts of law, which could lead to a different outcome. Thaler has expressed his intentions to appeal the decisions as he believes that DABUS is the rightful inventor. He opines that this may cause the courts to look at what manner and categories of patent eligibility will AI fit into and put implications to apply categories for AI to reduce ambiguity regarding the issue.

On August 27, 2019. The USPTO published an official notice in the federal register requesting the public to opine about the AI-related issues in the sphere of patents. This notice included questions regarding the sufficiency of the current patent laws and regulations pertaining to inventorship. The official notice questioned whether a revision of the patent laws and regulations is necessary to make the inclusion of inventions where an entity or entities other than a natural person contributed to an AI-related invention. But the Thaler decision shows that while USPTO may be open to hearing opinions, it is as yet far from having a positive view of AI-created inventions or from rocking the boat on patent eligibility issues.

Rahul is a seasoned IP Professional with 10 years of experience working closely with senior litigators on patent infringement and trade secret misappropriation. Rahul has a Bachelor's degree in Electrical Engineering from Indian Institute of Technology (IIT) Delhi and is a certified Project Management Professional (PMP). He has advised clients on more than 100 technology cases cumulatively resulting in over $1 billion in settlements and verdicts, including cases where he has testified at deposition or through expert reports.


Recent Insights
bottom of page