AI System Recognized as Inventor for the First Time, Named as Creator of a Food Container in Patent
The South African Patent Office made history in July 2021 when it granted a patent that names an artificial intelligence system as the inventor.
The patent covers a food container that uses fractal designs to create dimples and bumps on its sides. The new configuration, designed for the packaging industry, allows containers to fit more tightly together, simplifying their transport. The shape also makes it easier for robotic arms to grip the containers.
Patent owner and AI pioneer Stephen L. Thaler created the inventor, an AI system known as Dabus (Device for the Autonomous Bootstrapping of Unified Sentience). The patent's success in South Africa was achieved thanks to Thaler's attorney, Ryan Abbott.
Abbott and his team filed applications in 2018 and 2019 with 17 patent offices worldwide, including in the US, several European countries, China, Japan and India.
The European Patent Office (EPO), the UK Intellectual Property Office (UKIPO), the US Patent and Trademark Office (USPTO) and the Australian Intellectual Property (IP) Office rejected the application, but Abbott appealed. In August he won the appeal when the Federal Court of Australia ruled that an AI system can be an inventor under the country's Patents Act 1990.
The EPO Board of Appeal and the UK Court of Appeal recently ruled that only humans can be inventors. Abbott is asking the UK Supreme Court to allow him to challenge this point. He expects a decision to be made this year.
Abbott is a professor of law and medical sciences at the University of Surrey's law school. He is also a professor at the Geffen School of Medicine at the University of California, Los Angeles, and has written the book “The Reasonable Robot: Artificial Intelligence and the Law.”
He spoke about South Africa's decision during the virtual event “Artificial Intelligence and the Law,” held in September 2021 by the IEEE student branch at the University of South Florida in Tampa.
IEEE Spectrum interviewed Abbott to find out how an AI entity can invent something, the nuances of patent law, and the impact of the Australian and South African decisions on human inventors.
Inventing Systems
In 2014, Abbott began noticing that companies were increasingly using AI to perform various tasks, including design creation. According to him, a neural network-based system can be trained, for example, on data about different types of car suspensions. The network can then modify the training data, thereby creating new designs.
A second network, which he calls a critical neural network, can monitor and evaluate the output. If you tell the AI system how to evaluate new designs, and that you are looking for a car suspension that can reduce friction better than existing designs, it can alert you when a design matching this criterion appears.
“Sometimes AI automates the activity that makes a human an inventor on a patent. It occurred to me that these kinds of things will likely become much more common in the future and that they will have serious implications for research and development”
Some patent applicants have been told by their attorney to use a person's name in the patent, even if the invention belongs to a machine.
But Abbott considers this a “short-sighted approach.” If a lawsuit challenging the patent is filed, the registered inventor could be removed from their post during the court proceedings. If that person could not prove that he or she was the inventor, the patent could not be secured.
Abbott acknowledges that most patents are never challenged, but says it still concerns him. Meanwhile, he has found that companies using AI for inventions are increasingly worried.
“AI automates the activity that makes a human an inventor on a patent. It was unclear what would happen if a patent had no human inventor. There was no law on this anywhere. Just a lot of assumptions”
He and a group of patent lawyers decided to find a test case to help create legal precedent. They approached Thaler, the founder of Imagination Engines from St. Charles, Missouri. The company develops artificial neural network technology and related products and services.
Thaler created Dabus partly to generate and develop new ideas. He asked Dabus to generate an idea for a new type of food container, but did not tell the system what exactly to invent.
The lawyers determined that the food container design could be patented, as it met all the essential criteria: it was novel, non-obvious, useful and within the subject matter.
They first filed Thaler's application in the UK and Europe because, according to Abbott, these jurisdictions do not initially require filing an inventor registration. The patent offices conducted a standard review and found the application “substantially patentable based on the preliminary examination.” The lawyers then amended the application, naming Dabus as the inventor.
Typically, an inventor's employer becomes the patent owner. Although Dabus is not an employee, Abbott stated: “We argue that Dr. Thaler has the right to own the patents under general principles of property ownership, such as what is generally called accession, which refers to owning some property by virtue of owning some other property. If I have a fruit tree, I have the fruit from that tree. Or, if Dabus were a 3D printer and made a beverage container on a 3D printer, Thaler would be its owner.”
Impact on Inventors
Abbott believes that the decisions in Australia and South Africa will encourage people to create and use machines capable of generating inventive results and to use them in research and development. This, in turn, will promote the commercialization of new technologies.
He hopes the decisions will also encourage people to openly say whether their invention was developed by a machine: “The reason we have a patent system is to get people to disclose inventions, to add them to the public knowledge base in exchange for these exclusive rights.”
He says human inventors are likely to face more competition from AI in the future.
“AI has not yet reached the level where it will drive mass automation of research. When that happens, it will likely be in certain fields where AI has natural advantages, such as drug discovery and repurposing. In the medium term, human researchers will still have many ways to stay busy, while society enjoys significant advances in research”
