Posted on November 2, 2020
Authored by Aditya Krishna*
In the previous blog in the series on ‘AI and Authorship’, the author explored the discourse taken forward by the Tencent case and deliberated on its contributions towards the concept of Artificial Intelligence (“AI”) and authorship. This article will constitute the second blog of the series on ‘AI and Authorship’ and endeavours to analyse the current stance under Indian law with respect to the same. It highlights India’s approach to the concept of AI generated work and identifies the areas where the current laws are lacking. The article also identifies the current uncertainty and theorises the way forward by arguing for change in the judicial approach to solve the issues of AI and authorship.
Knowing Your Artificial Intelligence
Before getting into an analysis of the same it first becomes important to understand the distinctions in the types of AI that exist. For the purpose of this analysis we will be looking at the Type-1 distinctions which are based on capabilities of the AI systems. AI’s are accordingly categorised into three categories, namely (i) Weak/narrow AI, (ii) General/strong AI and (iii) Super AI.
- Weak/Narrow AI
These are the most common and simplest forms of AI. Such systems are goal oriented and are used to perform specific tasks. It usually functions within narrow constraints and does not in itself display intelligence similar to that of human intelligence. Instead, it tends to imitate human behaviour. They are used for a variety of tasks such as text-to-speech, voice assistance, internet browsing, online product recommendation etc. Such AI tend to require a heavy human involvement in its functioning.
- General/Strong AI
These are the more futuristic forms of AI which AI technology systems are beginning to move towards. It tends to exhibit human levels of intelligence and can understand and think in a similar manner to that of humans. It can also learn and perform tasks independent of human intervention. While currently, its applications remain theoretical, it is just a matter of time until AI reach this stage, especially considering most AI are already in the grey area between the weak and strong forms of AI.
- Super AI
The last type of AI, that is at this point is still very far away from being developed is that of super AI. Such systems have only been created in theory and are theorised to completely surpass human intelligence. For the purposes of copyright law and authorship though, in the opinion of the present author, the stance would remain the same as that would apply to strong AI, as both would work independent of human intervention.
An Analysis of Indian Copyright Law
Having discussed the three main types of AI, we can now move on to analyse the state of the current laws in our country. The main point of focus for the article will be the Copyright Act, 1957 (“Act”). It can be seen that the same principle laid down in the Tencent case is also enshrined in Section 2(d)(vi) of the Act which provides authorship of works created by computer systems to the individuals who caused the same to be created. Much like the stance followed by many other countries, India too mainly provides for authorship based on a human centric model of authorship. This can be seen through the various judgments of the Indian courts, where it was held that the human involvement is what creates the basis of copyrightability of the works so created. Another example of the same is the fact that the copyright office tends to provide authorship only to juristic or natural persons.
Having identified the importance given to “decisive” human intervention, it begs the question as to what extent the Indian law actually protects such AI generated content? From our understanding of the types of AI from the previous section, it can be inferred that the law could current protects only the simpler forms of weak AI that require a good amount of human intervention. This tends to be quite inadequate though especially in light of the current development in technology and the gradual movement towards strong AI. It remains to be seen how the courts will, in the future, interpret the aspect of human involvement to either include or exclude even such weak AI system, that require minimal human intervention, from its ambit. In the author’s perspective, the best way forward would be to interpret the human intervention requirement in the phrase “the person who causes the work to be created”, more liberally so as to cater for the diminishing importance of human involvement.
Linked to the previous question is the question of whether the current laws could possibly be used to credit authorship for human independent AI creations to the individuals involved? In the author’s opinion, the answer to the current question can be found by perusing through Section 17 of the Act, which deals with authorship rights of the first owner of the copyright. While stating the same, the author is fully aware that currently, the same cannot be used for the purposes of AI (mainly in light of the narrow interpretation of ‘Author’ under Indian copyright jurisprudence). Nonetheless, the author opines that the same still has the capability of being used in the future to secure such rights upon a shift in the stance on authorship and employment in the context of AI systems. While it has been argued that the provisions of Section 17 (a) or (b) cannot find applications due to the judicial interpretations of employment, the current author believes a more liberal interpretation of the same, when it comes to AI systems, could possibly be an easier way forward. Doing so would future proof the current laws and allow for their application even to human free AI systems used by individuals and companies. That being said, it is quite clear that the current stance merely caters to weak AI which involves human intervention and Section 17 provides no relief in the case of such AI-created works.
Overcoming the Paradox
As has been highlighted through this article, the current state of laws is such that it could either be used to provide authorship for human independent AI-created works or could be used to maintain the human centric authorship model. Much like Schrodinger’s cat, the exact outcome is yet unknown as we are yet to open the metaphorical box and see what factors could possibly play a role to lead to either of the outcomes. With that being said, the current author believes that a shift in the approach of judicial interpretation of the key terms identified above could possibly play a pivotal role in moving India’s copyright jurisprudence and laws into the future. Doing so would be convenient and also cater for the changing times.
While it can be seen that Indian laws are quite well situated, compared to that of many other jurisdictions, when it comes to AI-generated content, changes in the judicial approach could go a long way in future proofing our current laws. This requires a shift towards a non-traditional theoretical understanding of copyright law and authorship. Theorizing the same will form the main focus of the next piece of this series.
*Aditya Krishna is a third-year law student from Jindal Global Law School and is currently pursuing his B.A. LL.B. (Hons.) degree. He has a keen interest in Intellectual Property Law, Technology Law and Constitutional Law. He currently serves as a Contributing Editor at IntellecTech Law.