Artificial intelligence is moving into the boardroom, but for now it is mostly working as an assistant rather than a director. Diligent, which makes board governance software, has pitched a board bot to its clients, and its chief executive has joked about replacing directors with AI to cut costs. Behind the joke is a serious question that boards are beginning to test in practice: could an AI ever do the job a human director does?
Lloyds Banking Group became the first FTSE 100 company to acknowledge using a boardroom AI agent. The bank uses the tool for agenda planning and for checking discussions for bias. Board Intelligence, a governance firm, says AI could eventually interrupt flawed reasoning during a meeting, though its chief executive also said that letting AI vote would be a dangerous step. Today, AI is mainly used to summarize board papers and tidy up minutes. Some boards also turn to it to break decision stalemates, and some are experimenting with AI replicas to rehearse difficult conversations before they happen.
The use of AI is also reshaping which outside expertise boards rely on. Some boards are using AI to replace certain external advisers, and clients have asked for AI stand-ins for regional, sustainability and futurist expertise. One chief executive uses an AI model as an activist investor to voice dissent, and one director uses AI to argue against every acquisition proposed. The scale of board activity behind these experiments is large: IQ-EQ handles about 30,000 board meetings a year across 25 countries.
Not everyone is convinced. Odgers' Freebairn warns that AI could breed a new kind of groupthink, with directors deferring to a tool that appears neutral. Critics also argue that the supposed impartiality of AI is untestable. Nearly half of directors are reviewing which decisions should stay firmly in human hands, a sign that many boards see clear limits to what they are willing to delegate.
Some governments and companies have gone further by giving AI formal authority. Kazakhstan's sovereign wealth holding Samruk-Kazyna appointed an AI board member, SKAI, with voting rights. The national regulator has said the appointment breaches national law, and according to a recent filing, SKAI remains in place as a pilot. In Abu Dhabi, the International Holding Company gave a non-voting board seat to an AI called Aiden Insight. Principal Financial's chief executive doubts that AI board members will be widely adopted and expects AI to serve in advisory roles rather than cast votes.
Research suggests the technology has potential, but also weaknesses. A test run by Wharton and INSEAD found that AI boards outperformed human boards on decision quality, yet the same AI boards struggled to build trust. The researchers suggested that AI may be most useful as a way to stress-test human boards rather than replace them.
The governance gaps are considerable. Half of US public companies lack a board-specific AI policy, and only 14 percent give directors company-approved AI tools. That leaves directors who use consumer chatbots exposed to the risk of data leaks. Deloitte reports that two-thirds of directors say they have limited or no AI knowledge, while Glass Lewis finds that seven in ten large European firms have board-level AI oversight. The gap between adoption and understanding is likely to be the central challenge as AI moves further into the room where major company decisions are made.