Managing AI Is a Leadership Job, Not an IT Ticket

When an organisation adopts AI, it almost always files the decision under "technology." The IT team is asked to pick a tool, run a pilot, and report back. Leadership stays a comfortable arm's length away, treating AI as plumbing — important, but someone else's job. That instinct is the single most expensive mistake being made with AI right now.

Because the hard questions AI raises are not technical. They're questions of judgment, and judgment is what leaders are for. When an AI system makes a decision that affects a customer, an employee, or a citizen — who is accountable for it? What is the model allowed to decide on its own, and where must a human stay in the loop? What happens when it's confidently wrong? Whose data trained it, and would we be comfortable if that were public? None of these can be delegated to whoever configured the software. They are governance, and governance sits at the top.

This is what "AI management" actually means, and it's badly misnamed. It isn't managing the technology. It's managing the trust, the risk, and the accountability that the technology now carries on the organisation's behalf. A leader who can't say where their AI is allowed to act, who reviews it, and who answers when it errs has not adopted AI responsibly. They've outsourced consequential decisions to a system they don't oversee.

The organisations that get this right treat AI adoption the way they'd treat any transfer of authority — deliberately, with clear lines of accountability and someone senior owning the outcomes. The ones that get it wrong will discover, usually at the worst moment, that "the algorithm did it" is not a defence anyone accepts.

So the question every leadership team should be able to answer out loud:

When our AI makes a decision, who in this room is accountable for it?