Should AI Take a Board Seat? It Already Has. Nobody Appointed It.

The question arrives at every governance conference now: should boards appoint an AI director? The precedents get cited each time. A Hong Kong venture firm gave an algorithm a board vote as far back as 2014, and a Gulf conglomerate later installed an AI observer in its boardroom. The debate treats machine directorship as a decision still to be made.

The debate is looking in the wrong direction. In my work with boards across three continents, AI already holds the most influential seat at the table. It is simply not sitting in a chair. It sits inside the pre-read.

Trace the actual decision

Follow how a modern board decision is really formed. Management\'s papers are drafted with AI assistance. The market analysis is machine-synthesised. Directors prepare using AI summaries of packs too long to read in full, and the questions they carry into the room are increasingly machine-suggested. By the time the chair calls the item, the framing and often the recommendation itself have been shaped by systems that appear nowhere in the minutes. This is synthetic authority: influence without appointment or accountability. A human with that much sway over outcomes would be the most powerful person in the institution. This influence is invisible to the governance framework entirely.

I once reviewed a board pack for a chair based in Oslo who wanted a second read before a significant meeting. What struck me was not the quality of the papers, which was high, but the uniformity of their thinking. Every submission framed its problem the same way, reached for the same structure, and arrived at a similar shape of answer. The homogeneity was the fingerprint of a shared tool. The board believed it was receiving independent analysis from separate teams. It was, in a real sense, receiving variations on a single machine\'s worldview, and no one had decided that this should be so.

A board is a machine for questioning

The empirical warning is already on record. Wharton\'s cognitive surrender research shows that people adopt confident machine conclusions as their own without registering the transfer, and that the more polished the output, the less it gets questioned (Shaw and Nave, 2026). A board exists to question. An unexamined intelligence sitting inside its information supply defeats the mechanism at its root, because it shapes the options before the questioning even begins.

So my answer to the conference question is direct. Do not appoint an AI director, because appointment was never the issue. Recognition is. The Board AI Index exists to make this hidden influence visible, and most boards I assess sit at level two, Alert: aware the issue exists, blind to where it actually operates.

My recommendation is singular. Disclose the machine\'s role in the papers. Require every board submission to state where AI shaped the analysis or the recommendation, exactly as you would require a conflict declaration from a human adviser. What is declared can be interrogated. What is undeclared governs you.

References

• Shaw, D. and Nave, G. (2026). Thinking Fast, Slow, and Artificial: How AI is Reshaping Human Reasoning and the Rise of Cognitive Surrender. The Wharton School.
• Roos, J. (2026). The Missing Question in the AI Agenda. Global Focus, EFMD.
• Barker, K. (2026). Hidden Power: How Boards and CEOs Win the AI Era. Amplify.