ABSTRACT
Corporate laws like any other existing laws were formulated with the ultimate idea and quiet assumption that the chain of command is a human being. This is evidently highlighted under Section 166 of the Companies Act,2013 which enforce obligations on directors. Each duty presupposes a mind capable of intention, a will capable of corruption, and a person capable of being called to account. However, Artificial Intelligence is a threat to all the three concerns. In Indian boardrooms, the line between human judgment and machine logic has blurred. Algorithms have transitioned from silent assistants to the primary architects of corporate strategy, leaving directors to simply rubber-stamp complex decisions. This shift exposes a critical flaw in Section 166 of the Companies Act, 2013. This paper explores the “accountability vacuum” created when biased, opaque systems shape statutory disclosures and board deliberations. When AI obscures a material fact or distorts a risk assessment, the resulting legal harm is real, but the “responsible mind” required for fiduciary liability is nowhere to be found. By synthesizing global governance standards with India’s evolving regulatory landscape, this paper maps the “accountability vacuum” in algorithmic boardrooms and proposes a reform framework that mandates human oversight, algorithmic transparency, and specialized board literacy as prerequisites for fulfilling fiduciary duties.
Keywords: Section 166, Companies Act 2013, Fiduciary Duty, Artificial Intelligence, Algorithmic Governance, Corporate Accountability.
- INTRODUCTION
There is a scene, increasingly common in Indian corporate life, that the law has not yet learned to read. A board of directors convenes to approve a significant business decision an acquisition, a credit extension, a workforce restructuring. The recommendation before them is precise, data-rich, and confident. It has been generated by an algorithm. The directors deliberate and approve. The minutes record a human decision. The reasoning, however, belongs to a machine. Section 166 of the Companies Act, 2013 does not know what to do with this scene.
Corporate governance in India underwent a significant transformation with the enactment of the Companies Act, 2013, as it marked the first instance of formally codifying directors’ fiduciary duties under Section 166. director is required to act in “good faith” for the benefit of members, staff, shareholders, the community, and the environment, as provided under Section 166(2).[1] A director must use “due and reasonable care, skill and diligence” in addition to independent judgment, according to Section 166(3).[2]These duties are not procedural formalities. They are the law’s answer to a fundamental question: when a company is harmed by a decision made at the top, who is responsible, and why? Section 166’s answer is structural and deeply human responsibility attaches to the director because the director decided, because a human mind weighed, chose, and acted.
AI’s evolution from “advisory roles to proactive execution” marks a fundamental shift, with agentic AI systems now capable of autonomously pursuing long-term goals, making complex decisions, and orchestrating multi-stage workflows without continuous humanoversight.[3] Crucially, these systems do not arrive as neutral instruments. They carry within them the biases that are opaque, self-reinforcing, and capable of distorting deliberations, corrupting disclosures, and producing regulatory harm without leaving a traceable human fingerprint. The director who approves the output neither authored the reasoning nor, in most cases, can meaningfully interrogate it.
This paper terms the resulting legal condition the accountability vacuum, where fiduciary responsibility rests with a human director while substantive reasoning is performed by a machine lacking legal personhood and duties. Through a comparative analysis of Delaware’s Caremark doctrine, the UK Companies Act 2006, the EU AI Act, SEBI’s LODR Regulations, and the DPDP Act, 2023, it proposes reforms including a human oversight standard, algorithmic impact disclosures, and revised board composition requirements under Section 149.
[1]The Companies Act, 2013, § 166(2) (India).
[2]The Companies Act, 2013, § 166(3) (India).
[3]Mukherjee & Hannah Hanwen Chang, Agentic AI: Autonomy, Accountability, and the Algorithmic Society, arXiv:2502.00289v3 (Feb. 2025), https://arxiv.org/pdf/2502.00289v3.