Hey Alexa, Are You An Inventor? (Part 1)



Introduction

Artificial Intelligence (A.I) is the ability of machines, especially computers and computer systems, to perform tasks that are usually associated with the intellectual capacity of humans. Although the term was officially coined in 1956, attempts to unravel the possibilities of intelligent and self-learning machines began earlier. Since then, we have knowingly or unknowingly, welcomed A.I systems into our lives with open arms. The advent of virtual assistants has made us ditch our notepads and rely on Siri and Alexa to make lists and set reminders. Utter the incantation “Hey Google”, and get all the information you need within seconds. Other A.I wonders beyond voice assistants include text predictions, facial recognition, traffic and accident alerts, and chatbots. Needless to say, A.I is everywhere. Perhaps this is what led Sundar Pichai, the CEO of Google, to compare the profoundness of its impact with that of electricity and fire.

In its pursuit of reaching, or in some cases surpassing the levels of intellectual capacity of humans, A.I today has penetrated into areas which have traditionally been reserved for humans, like producing artworks and composing music. Going a step further, we now have inventions which have been generated by A.I. Here, it is pertinent to note the difference between ‘A.I-assisted works’ and ‘A.I generated works.’ The former refers to the application of A.I systems in the process of invention. In this case, the A.I system will be treated as any other machine or tool which is utilized in the inventive process. On the other hand, ‘A.I-generated works’ refers to the A.I system itself devising an invention with no human intervention. Conventionally speaking, shouldn’t the term ‘inventor’ be reserved for persons? Can the A.I system be called an inventor with respect to A.I-generated works? Keeping these questions in mind, let us analyse a recent claim for inventorship for an A.I system called DABUS.


What Is DABUS?

DABUS is an acronym for “Device for the Autonomous Bootstrapping of Unified Sentience” which is an A.I system created by Dr Stephen Thaler.


What Does It Do?

Artificial Inventors describes the system as follows: It is a “Creativity Machine,” which refers to a particular type of connectionist artificial intelligence. Such systems contain a first artificial neural network, made up of a series of smaller neural networks, that has been trained with general information from various knowledge domains. This first network generates novel ideas in response to self-perturbations of connection weights between neurons and component neural nets therein. A second “critic” artificial neural network monitors the first neural network for new ideas and identifies those ideas that are sufficiently novel compared to the machine’s pre-existing knowledge base.


The Patent Applications which listed DABUS as an inventor

In 2018 and 2019, Dr Thaler with the Artificial Inventor Project, filed patent applications in a number of countries through the Patent Co-operation Treaty for two inventions. One application, called ‘Neural Flame’ sought patent protection for a warning light for attracting enhanced attention. The other application called ‘Fractal Container’ was filed to patent an invention which is a new type of beverage container based on fractal geometry.


The Decisions by the Patent Offices

The United States Patent and Trademark Office (USPTO), the UK Intellectual Property Office (UKIPO) and the European Patent Office (EPO) refused the aforementioned patent applications. The common conclusion was that machines cannot be inventors.

According to the USPTO, a plain reading of the Patent Act shows the usage of terms such as ‘whoever’, ‘individual’, ‘person’, ‘himself/herself’ is a person. This implies that inventors ought to be natural persons. U.S caselaws provide that inventorship cannot be granted to even states or corporations, the reason being that conception is key to inventorship (Univ. of Utah v. Max-Planck-Gesellschaftzur Forderung der Wissenschaften E.V. : State cannot be an inventor, Beech Aircraft Corp. v. EDO Corp. : Inventorship different from corporate ownership, only natural persons can be inventors). Conception is understood to be a mental act as it refers to the “formation in the mind of the inventor, of a definite and permanent idea of the complete and operative invention.” Here again, the human element of inventorship comes into picture. Therefore, the USPTO refused the two patent applications.

In January 2020, the EPO published the grounds for refusal of the two patent applications. A.I systems are not considered as persons. Therefore, they do not have any rights, as for instance the right to be recognised as an inventor in patent applications. Therefore, DABUS cannot be an inventor. According to the European Patent Convention, if the applicant is not the inventor or a sole inventor, there must be a statement which indicates the origin of rights to the patent. Regarding this, Dr Thaler argued that since he is the owner of DABUS, he acquired the rights from the A.I system. However, the EPO cautioned that ownership must be distinguished from inventorship. The owner may own the output of the A.I system, but he is not the inventor.

Similarly, the UKIPO also concluded that DABUS is not a natural person and hence cannot be the inventor. Further, the A.I system did not have any rights that could be transferred to Dr Thaler. This led to an appeal at the UK Patents Court in September 2020, where the UKIPO’s decision was met with agreement.

The patent applications are pending in the Intellectual Property Offices of other countries. However, considering that patent legislations in most jurisdictions assume the inventor to be natural or legal persons, it is likely that the patent applications will get rejected in those countries as well.


Implications of granting inventorship to A.I systems

The Artificial Inventor Project claims that not recognising A.I systems as inventors and failure to give credit where it is due would devalue human inventorship. After all, who would like to give credit to someone who did not contribute to the work? As per this argument, if the A.I operator did nothing beyond feeding the information to the A.I system, they cannot be called an inventor as it is the machine that is doing the work. Therefore, while considering whether A.I systems can be inventors, it is crucial to keep in mind the degree of human intervention in the inventive process. Accepting A.I systems as the inventor may be equal to granting them the status of a ‘person’, which is unprecedented in itself. In an alternate scenario, can the patent application for A.I-generated works not mention an inventor at all? Conventionally speaking, a patent application must list the name of the inventor. In order to sidestep the problem of according personality to A.I systems to identify them as inventors, it has been suggested that the current system can be changed to include the possibility of not listing an inventor in such cases. While working on this alternative, its repercussions on the current incentive-based system must be considered.


Conclusion

It cannot be denied that A.I-generated inventions can disrupt the current patent system with the myriad of issues surrounding it. Although it is said that A.I technology has not yet reached a stage where it can generate outcomes with absolutely no human intervention, one cannot underestimate the pace of technological development. Thus, it is important for IP policymakers to consider the issues and weigh their outcomes accordingly.


References:



120 views1 comment