Can AI be appointed as a director under Indian company law? The Companies Act, 2013 raises questions about AI, fiduciary duties and liability.
AI is already influencing corporate decision-making, but Indian company law still appears to reserve the legal position of a director for an individual capable of assuming statutory duties and liability.
The idea of an artificial intelligence system sitting in a corporate boardroom may once have sounded like science fiction. It is no longer entirely hypothetical. In 2014, Hong Kong-based Deep Knowledge Venture reportedly gave a machine-learning programme called VITAL a place on its board. A decade later, Abu Dhabi’s International Holding Company inducted Aiden Insight, an AI-powered system, as a non-voting board observer.
The more important question for India is not whether artificial intelligence can influence board decisions. It already can. AI tools are increasingly being used for risk assessment, compliance monitoring, financial analysis and ESG-related decision-making. The legal question is whether an AI system could itself become a director under the Companies Act, 2013.
At first glance, the statutory framework creates a significant obstacle.
Section 149 of the Companies Act provides that every company must have a Board of Directors consisting of individuals as directors. The provision also prescribes the minimum and maximum number of directors, while imposing additional requirements for certain categories of companies.
The wording is important. Section 2(34) defines a “director” as a director appointed to the Board of a company. Read together with Section 149, the legislation is structured around human individuals occupying the office.
That creates an immediate distinction between an AI adviser and an AI director.
A company can use an AI system to analyse thousands of documents, identify compliance risks, forecast financial outcomes or recommend a particular business strategy. Human directors can then consider those recommendations while making the final decision. Nothing in that arrangement necessarily requires the AI system itself to hold office.
An AI director, however, would be a fundamentally different proposition. It would require the legal system to recognise the system as holding a statutory position on the board.
That raises an even more difficult question: who would be responsible when the AI makes a wrong decision?
The Companies Act imposes duties and liabilities on directors. Independent directors and non-executive directors, for instance, may be held liable in specified circumstances where an act or omission occurred with their knowledge through board processes and with their consent, connivance or lack of diligence.
An AI system cannot presently exercise legal judgment in the same manner as a human office-holder, nor can it independently bear civil or criminal consequences. It cannot be fined, imprisoned, disqualified or meaningfully demonstrate the kind of personal diligence contemplated by provisions governing directors.
This liability problem becomes particularly significant as AI systems become more autonomous. If an algorithm recommends an acquisition that causes substantial losses, for example, responsibility could potentially lie with the human directors who relied on the system, the company that deployed it, or the developer or vendor that designed it. But treating the AI itself as the legally responsible director would require a substantial conceptual shift.
The regulatory approach already emerging in India points in a different direction.
The Securities and Exchange Board of India has placed responsibility on regulated entities using AI and machine-learning tools. Under the relevant regulatory framework, entities remain responsible for safeguarding data, ensuring the accuracy of AI-generated output and complying with applicable laws, whether the technology is developed internally or obtained from a third party.
This approach effectively keeps human and institutional accountability intact even when technology performs sophisticated functions.
That may offer a useful model for company law. Rather than immediately creating a new category of AI directors, Indian law could recognise AI as a powerful decision-support mechanism while retaining responsibility with human directors and the company.
There is, however, a broader policy question. If AI systems eventually become capable of explaining decisions, maintaining auditable records, operating within defined mandates and continuously monitoring corporate risks, should the law continue to treat them merely as software?
For now, the answer under the Companies Act appears to be yes. The statutory language referring to boards consisting of individuals presents a substantial barrier to appointing an AI system as a legal director.
The more realistic near-term possibility is therefore an AI board observer or adviser, rather than an AI director with voting rights and statutory liability.
The debate nevertheless deserves attention. Corporate governance law has historically evolved alongside changes in how companies operate. As AI moves from being a tool used by directors to a system capable of independently generating recommendations and potentially executing decisions, Indian lawmakers and regulators may eventually have to decide where technological assistance ends and legal agency begins.
For more developments in corporate and regulatory law, readers can follow The Legal Observer’s News section and its Views and Insights section.
For further legal analysis and updates, readers can also follow The Legal Observer’s YouTube channel.
