India’s ‘Unregulated’ Tryst with Artificial Intelligence: Looking Into Future Without A Law?

Over the last few years, artificial intelligence (AI) has expanded enough to be the new normal in a country’s technological advancement. The advent of this technology varies from nation to nation and depends on various factors, including the economy, employment scenario, technology, literacy and over and above these factors, the outlook of the government. Remarkably, Sh. Arun Jaitley, in his last Budget speech as the Finance Minister, mandated the NITI Aayog to devise a national policy on artificial intelligence. Hon’ble Prime Minister has also emphasized upon the need for a solid artificial intelligence regime in the country. It is of little importance at this stage to highlight the importance as well as the need of this technology, which has been labelled as the most pathbreaking development in the past 100 years on the lines of electricity. The core area of this piece is the analysis of legal concerns surrounding this technology which is being accepted in a dangerously routine manner.

In pursuance of the direction issued by former Finance Minister Arun Jaitley, NITI Aayog has released a discussion paper recommending a national policy on AI. It also highlights the core areas for the application of this technology, which include healthcare, smart mobility and transport, education, agriculture, infrastructure and smart cities. The paper has extracted the core issues plaguing these sectors and subsequently proposed a methodology of curbing those problems using AI. But surprisingly, the paper conveniently chooses not to undergo a thorough analysis of the challenges surrounding this technology. Today, the debate is not around opting or non-opting of AI because I believe we have come past that. The debate is around regulating the import of this technology in the country within the realm of municipal and international law.

Data protection and AI

One of the foremost concerns surrounding AI is of data protection. After the series of incidents that unfolded with regard to Cambridge Analytica and involvement of the social media giant Facebook, the concerns have grown multi-fold. Every AI interface relies entirely on the data which is being fed into its system. Unlike other software, where human commands lead to results, AI systems develop their decision making based on a mechanical analysis of user data over the time. It is true that the very concept of artificial intelligence is the development of technology which can produce results falling parallel, if not better, than human outcomes.

Two major issues pop up at this stage:

1. Data protection framework in India;


2. In case of breach, who is to be blamed in the absence of human intervention?

The data protection regime in India is alarmingly weak to match the pace of growth of AI. Broadly speaking, the Information Technology Act, 2000 is the only piece of legislation which ‘touches’ upon this subject. Whereas it is undeniable that certain safeguards pertaining to data protection and privacy have been laid down in Sections 43A and 72 of the Act, but the safeguards fall greatly short of ensuring actual protection because of the obscure nature of provisions, added majorly through amendments. That the real mischief of the IT Act is not data protection or privacy is the fundamental flaw and raises the need for a comprehensive data protection legislation in India, on the lines of European Directive on Data Protection, UK Data Protection Act (1998), OECD Guidelines on the Protection of Privacy and Transborder Flows of Personal Data, 1980, and the Safe Harbour principles of the US.

Secondly, in the absence of the direct application of human mind behind any action that an AI system undertakes, who is to be blamed for the loss which is likely to be sustained by innocent users? With the outcome of the NITI Aayog paper, it is likely that the highlighted sectors may face greater involvement of AI in routine functioning of these sectors. The paper itself aims to tackle basic issues, including a shortage of doctors, automated trucking, high rate of drop-outs from school etc. Without questioning the efficacy of a computer system to deal with these issues, I believe the readers would definitely acknowledge the fact that this process involves micro-management of human data by a systematized machine, which may incur likely consequences of breach and misuse. Moreover, the classification made by NITI Aayog lacks homogeneity. It treats sectors like transport, smart cities and healthcare alike, which give rise to a glaring anomaly when analysed from the point of view of human behaviour. For example, according to PwC’s Report on AI in India, 93% of the users are hesitant in sharing their medical results despite knowing that this will enable them to have access to personalized health advice, but they may freely share their transport patterns.

Absence of a legal relationship

The next major concern affecting this advent is the absence of a legal relationship between humans and AI systems. Today, a vast majority of the users of advanced technology perceive these smart systems as a casual or friendly add-on to their experience. The simplest of examples include Apple’s Siri or Microsoft’s Cortana. If direct human-AI interaction is being contemplated by the government, would it be reasonable to leave the space for body corporates to claim “We didn’t do it, but our AI did” once legal injury ensues?

In the aftermath of the Supreme Court decision in Justice K S Puttaswamy (Retd.) & Anr. vs. Union of India and Ors., right to privacy has become operative as a duly enforceable fundamental right. However, in the absence of a legislation to deal with these upcoming challenges, it is more than likely that this right remains a piece of cryptic language for an average citizen losing his dignity in the digital space on an everyday basis.

 What is way ahead?

Isaac Asimov, in 1942, propounded the ‘three laws of robotics’ which greatly focussed on keeping a product of technology as subordinate to the human mind. On similar lines, first and foremost, it is crucial to ensure that an AI system is explicitly subject to all the laws of the land to which its human operator is subject. Secondly, the government must put in place adequate safeguards in the form of prior intimation of extraction of information to individual users i.e., to the source of information. Artificial intelligence is the future and there is no denial from the author, but in our quest of replacing ‘human errors’ with smart technology and strong AI, we need to move slow and keep abreast with the parallel needs of upgrading the laws and literary framework in the country.