Introduction:
Artificial Intelligence. Never have two words fanned the flames of curiosity of a generation of individuals since the birth of the Technological era. Artificial Intelligence, as the term itself suggests, refers to the constellation of technologies that enable machines to act with a level of higher intelligence emulating the human capabilities of sense, comprehension, computation and processed action[1]. Given the great innovative leap offered by Artificial Intelligence, it is rather surprising to hear about there being very few / virtually no laws regarding working, definition, boundaries and patenting of the same. In this Article, we shall deal with the contentious relation of Artificial Intelligence and Patent Law.
What is Artificial Intelligence?
Artificial Intelligence as a concept has been prevalent for centuries. Since the time of the Ancient Greeks and Romans when an “automaton” showcasing human tendencies was discussed to the era of explosion in Science innovation and fiction of the late 19th century early 20th century. However, the term Artificial Intelligence was first mentioned in the now landmark 1956 conference at Dartmouth College. It was here that the leading minds of America in the technological fields came to discuss, put forward and deliberate on Artificial Intelligence and its role in those times. Renowned British mathematician and cryptographer Alan Turing was one of the earliest proponents of Artificial Intelligence, even coming up with a test known as “Turing Test” to determine whether the test subject was human or Artificial Intelligence. Despite the human race having conceptualised the possibility of Artificial Intelligence ages ago, we still haven’t been able to properly define the exact boundaries of the term. This has been both a boon and a curse. It is a boon for it helps us grow and innovate every day with utmost freedom. It is a curse however as well for there lies no regulation or protection under the modern laws for the same, as seen specially in developing and under developed countries.
Patenting of Artificial Intelligence and Machine Learning:
If we are to momentarily don the hat of historians, we will notice a batch of interesting points in time period of the modern era. This is specially seen in the last century wherein the term “Patent” and “Intellectual Property” has gained significant traction. The time gap between the next great leap in the field of technology has begun to shrink rapidly. This can be seen especially in the area of Artificial Intelligence and Machine learning. Both these areas have seen one of the fastest growths, specifically in regards to their learning and use in developing nations in Asia and Africa. However, to protect such innovations, we have seen an advent in the areas of Patent Law, specifically in developed regions of the world. We can see the difference in development of patent protection of Artificial Intelligence as follows:
United States of America:
The United States of America is one of the leading developers in the field of Artificial Intelligence innovation and creation. Title 35 of the United States Code, Subject 101 (35 USC § 101) deals with the question of patentability. The wide scope offered was seen as a manner to boost business and technological growth. However, the Supreme Court of the United States (SCOTUS) in the landmark judgement of Alice Corporation Private Limited V/s CLS Bank International stated that patent claims directed towards subject matters pertaining to functions performed by any person showcasing an ordinary mental process or by a human being using a pen and paper and without showing any significant growth in the requisite “inventive step” would be denied. The two steps of Alice, despite being primarily for software, have been applied to Artificial Intelligence as well.
The above test was applied in the Blue Spike LLC v. Google Inc. and has subsequently lead to increased polarisation in the critical analysis of the Alice judgement. One of the biggest criticisms against the Alice judgement has been its oversimplification in deciding whether a particular subject must be patented or not. The two step process while seemingly made to ease the patent litigation only caused more confusion for it could be interpreted differently by separate courts. Moreover, the terms “software” and “Artificial Intelligence” are in no manner to be construed to mean the same merely because both rely on computer language and algorithmic functions. Artificial Intelligence is far more inherently complex in nature as compared to everyday software and thus must be analysed and determined on a case to case basis.
European Union:
The European Patent Office (EPO) is one of the first Patent Authorities of the world to delve extensively in the field of Artificial Intelligence and Deep learning. We see this in the guidelines regarding guidelines for Artificial Intelligence and machine learning. The EPO is one of the few offices to actively attempt to differentiate between mere algorithmic mathematical modules of everyday use and that used in Artificial Intelligence which is based on computational methods and algorithms regarding classification, clustering, regression and dimensionally reduction, such as neural networks, genetic algorithms and support vector machines among others. Guideline 3.3.1 is specific only to Artificial Intelligence and machine learning and lays down relatively modern grounds for patentability of Artificial Intelligence. In 2018, the EPO held its first conference on the patenting of Artificial Intelligence.
India:
In India the term Artificial Intelligence is yet to be fully comprehended and looked upon. Its use and bare purpose is known only to the upper echelons of urban India and there hasn’t been a pan India move to educate and inculcate the very use, purpose and nature of Artificial Intelligence in all strata of society. This limitation of knowledge has been extremely painful in growth and innovation in the field of Artificial Intelligence. However, with the advent of technology and internet connection to the rural masses there lies a hope that advancements in Artificial Intelligence will not remain mere conjectures of science fiction but rather everyday use.
The Patents Act, 1970 under Section 3(k) bars algorithms and programmes per say with no specification regarding the area of Artificial Intelligence and machine learning. Even the Computer Related inventions (CRI), 2017 does not delve extensively in the direction of Artificial Intelligence. The NITI Aayog in a Discussion Paper has bemoaned the lack of patent laws regarding Artificial Intelligence as a separate identity and not one limited to software, programs and algorithms. Currently, attempts to patent Artificial Intelligence are done in the same manner as seen with software in India, by attaching it to some form of novel hardware and put them up for patenting as one entity. However, should India truly feel the urgency to be in the race of technological innovation, they must update the patenting and patent examination grounds as seen in the European Union.
Conclusion:
The world of Artificial Intelligence is the world of the future. As human kind grows increasingly reliant on Artificial Intelligence, there is an ardent need for the laws of innovation and protection such as Patent Law to increase the specificity of manner, definition and boundaries of the same. Artificial Intelligence, while having the same base material as software, algorithms and software must be held as a different entity. To equate Algorithms, programmes and software of everyday use as seen under Patents Act, 1970 with Artificial Intelligence is akin to likening protozoa and amoeba with that of early human beings. There lies an extensive amount of difference between the two. Mayhap it is time the law recognised the same.
BY
SHREYAS SHETTY
BALLB-5th Year
ILS LAW COLLEGE, PUNE
[1] Page 12, NITI Aayog Discussion Paper on “National Strategy on Artificial Intelligence”