AI has quietly become the CEO's job. Nearly three-quarters of them just admitted it.

For most of the last decade, "our AI strategy" was a sentence a chief executive could say and then hand to the technology function. That era has just ended, and the people ending it are the chief executives themselves. The question is no longer whether AI belongs on your desk. It is whether you have accepted that it is now, unavoidably, part of the job.

The evidence is blunt. In BCG's 2026 AI Radar, a survey of executives across dozens of countries, 72% of CEOs say they are the main decision-maker on AI in their organisation, twice the share of a year earlier. The shift it describes is from a CIO-led initiative to a CEO-led one, and the logic is simple: AI has stopped being a tool the company adopts and become a way the company runs. When something touches strategy, operations, culture, risk and talent all at once, it cannot be delegated to any single function, because it is no longer a function. It is the business.

Why can't AI be delegated like other technology?

Because delegation works when a decision is contained, and AI is not contained. A new finance system stays largely inside finance. AI reaches into how you compete, how work is organised, what your people do all day, what you are liable for, and who you need to hire. Those are precisely the trade-offs a chief executive exists to make. Hand them down and you do not get delegation; you get a series of local optimisations that never add up to a direction, which is a large part of why most organisations fail at AI adoption. The same BCG survey found half of CEOs believe their own job security now depends on getting AI strategy right. When the person at the top has that much skin in it, ownership is not a choice; it is already true.

Doesn't "the CEO owns AI" just become theatre?

It can, and that is the failure worth naming. The unhelpful version of this news is the chief executive who starts performing AI: sitting in on model demos, learning to prompt for show, filling the calendar with visible engagement that builds nothing the organisation retains. Owning AI does not mean doing AI. The distinction matters, because a general-purpose technology does not reward the busiest leader; it rewards the one who builds institutional capability, which is the same reason your advantage comes from what your organisation knows, not the tools it buys. The job is not to become the most hands-on user in the building. It is to make the decisions only you can make and then leave.

So what does a CEO actually own here?

A short list of things that genuinely cannot be pushed down, and then the discipline to hold the organisation to them.

1. The direction. What AI is for in your business, and just as importantly what it is not for. A national-scale technology needs a company-scale decision about where it earns its place.
2. The ownership. Who is accountable for AI outcomes across the business, so it is not everyone's job and therefore no one's. These are the questions a board should be asking.
3. The standard. What good and safe look like, decided once and clearly, so thousands of downstream choices have something to align to.
4. The capability, not the theatre. Invest in the skills, data and governance that outlast your attention, rather than your personal hours at the keyboard.
5. The willingness to choose. AI forces genuine trade-offs between speed and control, cost and risk. Naming them is the chief executive's work; avoiding them is how the strategy quietly stalls.

What does this change for me as a leader?

It changes what you can honestly outsource. You can delegate the rollout, the tooling, the pilots and the training. You cannot delegate the decision about what AI is for and what you are prepared to trade to get it, because that decision is the strategy, and the strategy is yours. The chief executives now claiming AI as their own are not doing so because it is fashionable; they are doing so because the alternative, a company running an accidental AI strategy assembled from a dozen unaligned local calls, is worse.

This is the real content of the end of business as usual: the technology has climbed high enough up the organisation that it now sits where you sit. Take the few decisions only you can take, build the capability that survives your attention, and let everyone else get on with the doing. That is ownership without theatre, and it is the version that actually compounds.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.