Can An Invention By Artificial Intelligence Be Patented?

A powerful antibiotic that is effective against numerous infections was developed in 2020 with the aid of a machine-learning system. AI is also being utilised to advance space technology, ship design, drug discovery, vaccine development, and vaccine administration. There could be a lot of AI-related inventions in the next few years and this poses one of the greatest dangers to patent systems ever.

Patent language specifies inventions in terms of what does what, why, when, and how, therefore it is difficult to patent AI. In other words, you must give a structural description of your AI invention. However, courts do not view computer components like processors and computer memory as machine structure since they believe that they just implement functions and do not carry out any product transformations. As such, many machine learning models that are applying for patents are at risk.

What is patentable?

Generally, an invention cannot be patented unless it satisfies all of the following conditions. Firstly, there must be one or more inventors for an invention to exist. This covers practically all technologically related goods, procedures, or techniques. Additionally, the invention must be new, which implies that it has not yet been commercialised. Your innovation’s uniqueness is confirmed through the patent process, but you must convince the patent examiner of this. Novelty is a component of the patent examiners’ review process.They contrast a patent application’s claims with those of prior patents.

Next, the aspect in question is whether the concept is original and not obvious to someone skilled in the art. What does ‘someone skilled in the art’ mean? The patent examiner is looking to see if this invention is trivial, would it appear to be obvious to anyone with an average level of training in the field or professional knowledge. In addition, a patent must be capable of industrial application or utility. The invention can be made or used in industry, as is claimed. To be useful, the invention must have a purpose and/or has economic significance.

DABUS and Cases of AI Patent

The US Patent and Trademark Office (USPTO) denied a patent petition filed on behalf of DABUS, an artificial intelligence (AI) system in 2020. A food container with a fractal surface that aids in insulation and stacking, as well as a flashing light for calling attention in an emergency, were the products DABUS requested patents for. They were submitted by physicist Stephen Thaler, the creator of Imagination Engines, a business that conducts research and creates artificial neural networks like DABUS.

The schematics for a shape-changing drink container invented by AI system DABUS.
Source:BusinessInsider

DABUS, according to Thaler’s petition, was a “creativity machine” that could recognise the “novelty and salience of the instant innovation.” However, considering earlier federal decisions on the terminology pertaining to the nature of invention, the USPTO determined in its final opinion that “patent laws require that an inventor be a natural person.”

Earlier, in a parallel case involving a copyright issue with Thaler’s AI system, a US federal circuit court upheld a 2021 decision confirming that, as per the language of the Patent Act, AI systems cannot patent inventions because they are not human beings. With his counsel criticising the circuit court’s “limited and textualist approach” to the Patent Act, Thaler intends to appeal the decision.

The DABUS decision profoundly alters how intellectual property (IP) is viewed, and it makes clear a problem that patent systems are just now beginning to face: Can an AI-based invention be granted a patent? The DABUS case, according to Toby Walsh, Laureate Fellow and Scientia Professor of AI at UNSW Sydney, demonstrates how AI will more frequently challenge intellectual property law and be used to support invention.

The DABUS case also marks the first instance in which an AI system has been named the sole inventor, posing a conundrum for legislators throughout the world as they attempt to unify international legal views regarding AI patent law. For similar reasons relating to personality, patents for DABUS were also rejected in the UK, Europe, and Australia. The European Patent Office questioned who would enforce the rights granted to an inventor in such a situation.

While the DABUS case represents the first instance in which an AI system has been recognised as an inventor, AI has played a role in innovation before. A deep learning algorithm was used in 2019 to find a chemical compound effective against drug-resistant bacterial strains and create an antibiotic called Halicin. It was chosen from a collection of more than 100 million molecules that combat different pathogens.

According to Professor Walsh in a press release, Halicin was initially developed to treat diabetes, but its efficiency as an antibiotic was not known until artificial intelligence (AI) was tasked with looking through a sizable database of medications that could be converted into antibiotics. Therefore, there is a mixture of human and machine in this discovery.

Halicin was a compound identified by the machine learning software.
Source:ChemistryWorld

Professor Walsh further highlighted some of the challenges with ownership regarding AI-related inventions. There are a bunch of fundamental problems such as the ownership of the IP as it cannot be the AI. Does AI change the nature of invention? According to current laws, invention must not be obvious to an expert capable in the state of the art.

Limitations of Patent Law

Since the foundation of patent law is the idea that inventors are people, dealing with an inventor who is a machine is currently difficult. As patent applications naming an AI system as the inventor have been filed in more than 100 countries, courts all over the world are currently debating how to handle this issue.

Public consultations on AI and intellectual property (IP) law are being held by various organisations in the US, UK, and Europe. The WTO’s 1994 Agreement on Trade-Related Aspects of Intellectual Property Rights established certain requirements for patentability of inventions (TRIPS). While an AI may be significantly more capable, will patent offices be able to keep up with its speed?

According to Associate Professor Alexandra George, an expert in IP law, if we accept that an AI system is the true inventor, ownership will be the first significant issue. Who actually owns the property is unclear. An AI is not recognised as a legal person, which is a requirement for an owner. Professor George and Professor Walsh, who co-authored an article for the scientific journal Nature, argue that governments all over the world will need extensive legal reform to decide whether to grant IP protection to AI systems. They suggest creating a special type of IP law called AI-IP that would be designed to safeguard the inventiveness of AI.

A distinct AI-IP doctrine has the advantage that it could be customised to meet the unique circumstances in which AI creativity occurs, as opposed to retrofitting outdated patent laws to account for new technologies like AI. Unlike patents, which are typically given to the inventor, lawmakers may decide to divide the rewards from an invention produced by artificial intelligence (AI) differently, perhaps between the person who developed the AI, the person who is directing it, and the owner of the data that was used to train it.

An international treaty governing AI patents should also be formed, much like what the TRIPS agreement did in governing the use of designs, trademarks, copyrights, and other areas of IP. It would set out uniform principles to protect AI-generated inventions in multiple jurisdictions. If courts and governments decide that AI-made inventions cannot be patented, the implications could be huge. Funders and businesses would be less incentivized to pursue useful research using AI inventors when a return on their investment could be limited. Society could miss out on the development of worthwhile and life-saving inventions.

Companies with the most AI related patents/
Source: EEWorldOnline

In short

If a legal framework for patenting AI-generated inventions is not achieved, there will be implications in attracting investment in cutting-edge industries. That does not mean you should give up on patenting your AI. Rather, it’s important to understand how your AI will be approached from a patenting perspective. If you ever need assistance, reach out to us at Intellect Worldwide as we are the leading experts in patenting.