According to the World Intellectual Property Organization (WIPO), the AI technique showed an average growth rate of 28% between 2013 and 2016.
From 1956, which was the year that AI was invented, to 2017, more than 1.6 million scientific articles related to AI have been published. In this same period, approximately 340,000 patent applications of AI-related inventions have been presented.
In 2011, 12,473 AI patent applications were filed; at the end of 2017, the WIPO registered 55,660 AI patent applications, a patents’ increase of approximately 300% in this period.
Can what AI creates fall within the definition of an invention for the purposes of law? How much of what AI creates can be attributed to AI completely? Can any of it at all? Will the data sets used to train AI be considered part of the invention? Is AI without human intervention capable of inventing? At this point, there are more questions than answers, and naturally so.
In 1956, at a conference in Dartmouth College, the term Artificial Intelligence was coined. It originated as a pursuit of creating true, humanlike intelligence. Today, it has revolutionized the world and what we do. But has it achieved its ‘true purpose’ – of mimicking human intelligence? As of today, it doesn’t seem so. But it’s surely getting there. By extending inventorship to AI, what we are aiming to do is more than merely have it mimic human intelligence. We’re (with the way the laws today are) looking to have it mimic a human being in its entirety.
Stephen Thaler filed simultaneous applications for grant of two patents (one for a flashlight, the other for a food container) to an AI machine DABUS (short for Device for the Autonomous Bootstrapping of Unified Sentience). All three major patent offices – the USPTO, EPO and UKIPO rejected the application. The law requires the inventor to be a natural person. The reason for such a requirement is simple. It stems from the fact that a patent comes with enforceable rights. It takes a natural person to enforce those rights.
Academicians and attorneys seem to have diametrically opposite views on the position of AI as an inventor. Academicians are more hopeful, for lack of a better word, that AI be treated as an inventor and be granted patents. The legal fraternity thinks that the law needs an overhaul if we were to have AI as “inventors”.
Determining non-obviousness is essential in the patent process. These evaluations are performed by people skilled in the field of the “art” to which the invention belongs. So, when an inventor is AI, we’re going to need a new set of “AI skilled in the art”. Yes? Yes.
Machine Learning is a huge component of AI. Algorithms are trained with previously generated data, with which the algorithm learns skills and parameters that it applies to new data to perform its functions. The performance of the algorithm/AI thus changes with the training data that it is fed. Which means, the AI is constantly evolving. The thumb rule of patents is full disclosure in the patent application. With a constantly evolving algorithm, that becomes impossible. Will the disclosure of the initial algorithm ever be sufficient for the full disclosure of the invention?
Assuming we overcome challenges posed to laws in the patent application process vis a vis AI, what happens if a patent is successfully granted? The purpose of IP is to promote innovation by creating an atmosphere where such innovation is rewarded. One such mechanism is licensing and transfer of IP. The law is designed for such transactions to be carried out by legal persons, simply put, human beings. The challenges that AI authorship pose thus go beyond an IP law overhaul but extend to a change in laws on a much larger scale.
A report on AI and IP Policy released by the USPTO details the lacuna in the law as is, and possible changes. The Report acknowledges that US patent law is replete with language indicating that an inventor must be a natural person. The report was opened up to comments and from the responses received, the following could be ascertained:
The assumption is that AI technology is incapable of conceiving an innovation by itself. The report goes on to state that “the use of an AI system as a tool would not preclude the natural person from qualifying as an inventor (or joint inventor). While the possibility of AI ownership in the future is examined, it seems that most people believe that Artificial General Intelligence that needs to be reached for AI to think and invent on its own is in the distant future. Consequently, the consensus reached was that there is no immediate need to change the law.
One major concern was the threshold of non-obviousness that people believed AI had the potential to affect. Accordingly the same needs modification to cater to AI intelligence beyond the “ordinary skilled artisan” threshold.
The EPO is of the same view as the USPTO in terms of the main function of intellectual property rights being to protect the interest of human creators and encourage innovation. Certain AI-generated works, the EPO believes may be construed as “intellectual works” and be capable of copyright with the ownership assigned to an ultimate human creator.
The European Union is coming up with an AI Regulation that proposes to regulate AI use cases. The draft rules propose fines of up to 4% of global annual turnover (or €20M, if greater) for a set of prohibited use-cases.
China is one of the biggest players in the AI arena. In July 2017, China’s State Council released the country’s strategy for developing AI titled New Generation Artificial Intelligence Development Plan that outlined China’s aim to be a world leader in AI by 2030. China aims to monetise AI into a trillion-yuan industry, and to emerge as the driving force in defining ethical norms & standards for AI.
We are all trying to keep up with AI advancements, while being responsible for the same. What a time to be alive! We are also at a stage where we really do have more questions than answers. Will the law be able to keep up with AI advancements? Only time will tell!