October 20, 2019
Right to Privacy in AI era?
October 20, 2019


The reality of intelligent machines seemed too far-fetched just a couple of decades ago. This idea lived a purely fiction existence in science-fiction novels and movies that largely portrayed the machine taking over human lives and posing a significant threat to the “most intelligent” being on the planet (cue- Terminator, Robocop). Thankfully this fiction is not on its way towards actualization. Yet, the growth story of Artificial Intelligence from being identified as just an abstract idea in 1974 to a concrete reality that is moving towards becoming a very integral part of our everyday lives must push us towards thinking of its policy and legal implications.  
With the increasing usage of AI, it has become evident that the law has not yet caught up with the changing technology. With AI systems becoming smarter with each passing day, the possibility of the AI inventor coming into the picture is becoming increasingly plausible. The disruption that AI would create in law’s empire and particularly in the domain of IPR can be identified in the fact that it fundamentally questions the concept of agency, authorship, creation and originality. The upheaval that AI would create can also be understood in terms of the liability accrued to it. While the question of AI agency can be answered at a time when the reality of artificial humans comes closer, this post intends to look at the existing framework in copyright and patent law in India and check the possible tangents it could take to accommodate the raise of AI.  
Why this issue now?
The machine inventor is not actually a recent development; it is rather the creation of Machine Learning software (ML) that has brought with it, questions on inventorship. While traditionally machines have aided creation, it is only now that we are seeing machines creating on their own products. ML has created the capacity for machines to process data in a manner that the outcomes are intelligent and capable of comparison with human cognitive processes. With theinput fed to it, ML uses an in-built algorithmic response to learn from the data.  
Inventions by automated machines currently do not have any protection. While several jurisdictions outrightly deny copyright and patent protection to non-human inventions, in other countries such as the UK and India a route for possibly extending ownership to programmers who created the AI machine can be found. This connection is drawn based on their usage of the term ‘author’ in their copyright statutes to refer to the creator. This, however, still comes back to the human agent.  
To understand why AI’s impact on IPR needs to be seriously considered, the following hypothetical scenarios that range from existing AI applications to possible innovations, should be examined to see the inadequacy of the legal provisions.  
The scenarios are as follows –  
  1. Consider an intelligent movie-making application that allows a user to compile videos through their input but modifies the same through its own internal algorithm to improve the quality of the same – supposing this leads to a brilliant movie-like output which the user may want to copyright, does the copyright over the same lie with the user, the creator of the programme or the intelligent application? 
  2. A Machine Learning AI like Google’s Deep Mind which produces news articles, in its process of learning, infringes copyrighted work. Whose liability would the infringement be – Deep Mindor the specific programmer handling the AI?
  3. With the rate of innovation, it is foreseeable that AI might also be able to make a claim for agency and subsequently ownership over the works it produces – at a situation like this would the law be willing to accommodate non-human creators? 
The existing framework  
Section 6 of the Indian Patents Act, 1970 identifies that only ‘persons’ can apply for patents in the Indian regime. The Copyright Act of 1957 also defines authorship at various places purely in terms of the human agency. The attachment of human agency to the IPR regime can be understood given that intellectual labour finds itself to be at the center of the reasoning for according protection to creative and inventive works. Legal protection is justified for this intellectual labour as this would foster an environment that works for the promotion of human welfare through incentivizing innovation. However, with AI, there is no metric for judging intellectual labour. Although the process of ML does mimic human cognitive processes, there remains an apprehension with calling this intellectual labour because of the lack of preconception. The inventive product of AI is a result of the working of its algorithms rather than it choosing to create. Therefore, AI creations cannot be said to fall within the identified justifications for according it protection. Further, AI does not require the same motivations as humans do for creating.
While looking at the 3 mentioned scenarios from this existing framework, the possible solutions for same would be –
  1. The copyright for the output would lie without the user as they are the most proximate creator. The third possibility of AI owning the copyright is explicitly ruled out. Between the user and the programmer, the user would be a more appropriate owner as under Section 2(uu) of the Copyright Act as they would be the producer (Section 2(d)(v)).
  2. This scenario presents a tougher question. Copyright infringement as identified under Section 51 the Indian Act says that a “person” is an infringing agent. Therefore, there is no scope of accruing liability on non-human actors here, once again eliminating AI from the picture. The programmer would be liable although they did not have a role to play in the infringement.
  3. The scenario presents the most complex situation that fundamentally questions the definition of personhood and agency. This would not only disrupt the IP regime but also force us to think of new possibilities for rights and liabilities.
Patent Law also would face disruption with the raise of AI as the framework of it world over was established at a time when this technology seemed impossible in the near future. Right now, there are more than 30,000 AI-related patents that have been filed by Google and Microsoft alone. With big tech companies embracing AI, it can easily be foreseen that AI inventions will be a part of the Indian fora.  
 The AI inventions patented in the US have been successful because a human inventor is identified in the patent application. Patents without the identification of a human agent have been rejected in Canada, Australia and by the European Court of Justice. A common thread of arguments that can be identified in the reasoning given is that the recognition of the AI inventor takes away from the aspects of originality and preconception that is associated with an innovation. The introduction of the patent regime was largely justified as being the system that was needed to be in place to promote innovation and invention by human creators by incentivizing the same. A connected argument arising in the same context is that recognizing an alternate form of intelligence could lead to the atrophy of the inventive human mind.  
Where we need to go?
At present, we stand at the right juncture to decide how we would like to design our legal regime to determine the answers to these questions. Instead of the law playing the game of catch-up, it would be in the best interests of all stakeholders involved to begin thinking of legalistic changes from now.




Namratha Murugeshan is an undergraduate candidate of Law (B.A.LLB Hons.) at NALSAR University of Law, Hyderabad.



In content picture credit: Moneycontrol

Leave a Reply

Your email address will not be published. Required fields are marked *