The phrase that is de rigueur in today’s corporate world is Artificial Intelligence (AI) and that is not without reasons. It is changing the scenario completely – automatic product manufacturing, to strategic initiatives, to processing voluminous data, to literally drafting many a legal document.
True, today data is information and anyone well-equipped in that area – be it an individual or a company (big or small) will impact the target-audience and market considerably. A company will be able to foresee risks, probably prevent frauds and maintain operational efficiency, with the robotic tools at its disposal. Human behaviour or a person’s predilections too are within striking distance. A simple example – I love watching thrillers and every time I finish a movie, it prompts to the next one – saves my time and keeps my interest in that platform or app sustained. For a company, investor satisfaction is key, and most FAQs are imbued with answers thereby ensuring that the Investor Query is resolved saving once again human time.
However, too much dependence on algorithms (which run the AI) can be detrimental in the sense that inherent biases fed into the data can lead to unpredictable outcomes. A human eye and an innate ethical consideration that a human being is equipped with can never be replaced.
The question that now arises is that, does India have sufficient legal protection and, if not, are there any such endeavours on the horizon?
The Companies Act 2013 (the ‘Act’), as amended, does not address AI issues, and therefore, there isn’t any protection for any data spill or a perverse analysis. The managing director/CEO with his/her board of independent directors is in a fiduciary capacity towards the shareholders and owes a duty of reasonable care to the customers of the company. Apart from watching the back of each employee, he/she can be held accountable to the public at large if the company is listed. Has the director done the due diligence before relying upon the processed data or has he/she been remiss in blindly relying upon it. Has independent judgement been exercised – unaffected by what the AI unit has laid in front of him/her? In essence, as per Section 166 of the Act, he/she has acted in good faith, avoided conflict of interest and undue gain and exercised due care.
One of the biggest risks one will have to keep in mind is the transparency factor. The company has to choose between white-box and black-box algorithms. White Box are simpler ones, with easy-to-understand processes. But, at times, if the data is complex, a black box one is preferred. But it can mask the methodology, as it arrives at decisions – say for example – calculating the insurance premium or a sanction of a loan or sifting out suitable recruits – all of which can affect the future of the company. Shifting responsibility is an absolute bar because the director or the key managerial personnel is/are ultimately responsible for any discrimination – a commission or an omission. Most companies have now started training directors in this area so that their schooling eventually eliminates any biases and complies with all applicable laws including anti-discrimination laws, which AI can infract by processing an inherently biased given data.
Much hue and cry has been raised about the ‘right to privacy’. Naturally, AI will not be responsible for any data leak that is already out in the public domain. The Digital Personal Data Protection Act (DPDPA), 2023, has a framework but it is insufficient to marry AI with robust corporate governance. What is required is a clear-cut stipulation of AI provisions spelling out discrepancies, lapses, violations and the requisite punishment.
A bizarre theory was floated that there perhaps could be ‘digital directors’! Can one hold them legally responsible and even if one could can you punish them – at most, you could fine the company. An interesting case had come up in the Delhi High Court in an Intellectual Property case – Christian Louboutin Sas and Anr vs The Shoe Boutique-Shutiq, where the court held that “The responses and inputs of ChatGPT cannot be claimed as a basis of adjudication of legal or factual issues in the Hon’ble Courts….. Furthermore, there are possibilities of incorrect responses, fictional case laws, imaginative data etc. generated by AI chatbots. Accuracy and reliability of AI generated data are still in the grey area.” When a data itself cannot be relied upon, there is little scope for any concept of a ‘digital director’ with any attendant fiduciary duty of care.
The government of India, through the ministry for electronics & information technology (MeitY), had clarified last year that the government would regulate AI to the extent of protecting users from harm although a specific legislation may not see the light of the day soon. It had acknowledged that an AI framework for spurring the growth of emergent technologies was incumbent to promote fairness, transparency, remove uncertainties, standardise a procedure and promote best practices that are both within the ambit of the law and ethics.
MeitY had, a decade ago, constituted seven expert groups for designing India’s AI programme. On 13 October 2013, these bodies tendered the first edition of a formal report (the ‘AI Report’). The AI Report is designed to become a roadmap for the development of India’s AI ecosystem factoring in governance, transparency and intellectual property with an adequate legislation and ethics firmly in place to ensure the collection, storage and processing data following the highest standards of excellence.
Chairing the Global Partnership on Artificial Intelligence (GPAI), India also hosted a three-day summit (the ‘GPAI Summit’) in New Delhi in December 2023. The GPAI Summit saw 29 GPAI members reach a ministerial declaration to advance a safe AI by developing regulations and policies. Earlier, in December 2023, the ministry for external affairs hosted the Global Technology Summit for the purpose of deliberating upon the transforming nature of technology and geopolitics.
These steps do show a serious thought being given by the government to AI but it cannot substitute a strong legal and ethical framework thereby ensuring a solid corporate governance. The ‘environment and climate change’ have occupied centre-stage, but AI needs to be forming the cornerstone of any corporate governance reform, as it is here to stay.
AI presents a revolutionary opportunity for Indian companies. Responsible AI governance practices, replete with legal and ethical considerations, can foster a collaborative approach, engaging the government, industry and public at large, to promote efficiency and productivity and put India on the global map.