The history of patent filing dates back to the 15th century, with countless patented inventions being recorded ever since. For over 500 years, only humans have been creating innovations and receiving patents.
At first glance, it makes sense — who else would file for patents? Well, some computers and robots have also started to invent things using AI, but the current patent laws prohibit these machines from getting the patent for their works. So now what can the machines do? And most importantly, should they get to exercise ownership of their inventions and have intellectual rights like humans?
The rise of advanced artificial intelligence-based machines is posing new and unexpected challenges to the current patent laws followed in most countries today. The latest case highlighting the problem is related to DABUS, an AI-powered system developed by material scientist Stephen L. Thaler’s AI company (Imagination Engines). DABUS is considered the inventor of an enhanced food container and an autonomous bootstrapping device by its creators.
But DABUS can’t own patent rights.
Stephen Thaler filed patent applications on behalf of DABUS in the US, UK, Australia, New Zealand, and at the European Patent Office (EPA) so that his machine could receive legal intellectual proprietorship over his inventions.
But everywhere he tried, the concerned authorities declined to award patents to the machine. Thaler then appealed in federal and high courts in most of these countries, but many of his appeals are either pending or have been rejected by the judges stating that law does not consider AI an inventor. Thaler’s appeals seem unlikely to succeed, which seems rather unfair: why shouldn’t the rightful inventor be allowed to own the rights? Well, it’s complicated.
Old patent laws won’t work for AI-based machines
A recent commentary published by researchers from the University of New South Wales (UNSW) in Sydney, Australia sheds light on the different aspects of the AI vs patent law conflict, as seen in the DABUS case. According to the researchers at UNSW, the authorities and courts refused to give a patent to a machine because, in the eyes of the law, artificial intelligence is not a legal entity.
The patent laws allow only human beings or entities comprising humans to be considered as creators, owners, innovators, and patent holders. Therefore, intellectual property (IP) rights such as patents can only be awarded to a human and not a machine. Explaining this further, one of the commentators and associate professor at UNSW Alexandra George said:
“Even if we do accept that an AI system is the true inventor, the first big problem is ownership. How do you work out who the owner is? An owner needs to be a legal person, and an AI is not recognized as a legal person.”
Suppose the law permits AI to receive patents then another conflict arises. IP rights are given on the basis of ownership, so if an AI invents something then who would be considered the owner — the machine? Or the company or inventor of the machine? Or the software on which the machine runs? Or the firm or user who bought or currently owns the machine? Or the programmers who wrote code for the machine using which the machine became an inventor? The loops only become more complicated when you consider all the implications that arise from granting patents right to an AI; and yet, this seems to be a shortcoming of our legal system as much as an ethical conundrum.
The researchers argue that the current patent law would not be able to give answers to such questions concerning ownership in the case of AI. This is also why they believe that since DABUS is not a legal entity, it would be very challenging to have it secure a patent in accordance with the existing laws.
So is there any solution to this AI-patent law conflict?
To deal with cases like the one that involves the AI system DABUS, the existing patent and IP law needs to be amended, upgraded, and modernized. However, that does not imply that the current laws are unjust. Second commentator and AI professor at UNSW, Toby Walsh points out that patents are given for inventions that are novel and not obvious.
Whereas machines innovate or create things by storing and processing all the obvious information that humans provided them with. They don’t have any novel ideas or thoughts of their own. In an interview with Tech Xplore, Professor Walsh explained:
“There are certain assumptions built into the law that an invention should not be obvious to a knowledgeable person in the field. Well, what might be obvious to an AI won’t be obvious to a human because AI might have ingested all the human knowledge on this topic, way more than a human could, so the nature of what is obvious changes.”
The researchers further suggest that the conflict can be resolved by introducing “sui generis” or “AI-IP” in the existing IP laws. This means that the rules and regulations concerning patents should be modified in a way that they could consider AI-generated inventiveness in a reasonable manner. The researchers are now looking for answers related to the different possibilities in which an AI can emerge as an inventor.
With AI becoming more and more refined, the pressure will only grow for a solution. For now, it’s not clear what such a solution would look like.
The commentary is published in the journal Nature.
Was this helpful?