Corporate governance has always assumed that a natural person stands behind every decision a company makes, someone who can deliberate, be persuaded, and ultimately be held to account. Boards now routinely delegate substantive parts of that deliberation to algorithmic systems: AI tools screen credit risk, flag related-party transactions, model scenario outcomes for capital allocation, and, in a small but growing number of firms, sit in the boardroom itself as non-voting observers. This paper examines whether the fiduciary architecture of the Companies Act, 2013 built around Sections 149, 166 and 134, and premised on the natural-person director as the locus of judgment and accountability can absorb this shift without dilution of substance. It surveys the doctrinal debate on AI legal personhood and the case for and against a Board Service Provider model, examines the Securities and Exchange Board of India’s new Regulation 16C regime assigning regulated entities sole responsibility for AI/ML tool outputs as a template for board-level accountability, and situates the Indian position against comparable developments in Delaware oversight jurisprudence and the EU AI Act’s obligations for high-risk systems. It argues that the central risk is not that AI will formally displace directors, but that the duty of independent judgment under Section 166(3) will be hollowed out through uncritical reliance on algorithmic recommendations a phenomenon this paper terms ‘rubber-stamp liability’. It proposes a governance framework built around mandatory algorithmic-decision registers, a heightened ‘meaningful contestability’ standard for director reliance on AI outputs, board-level AI risk committees for companies above a materiality threshold, and statutory clarification that delegation to an algorithmic system does not, without more, discharge the duty of care under Section 166(3).