Why is AI the thing keeping you up at night?

It is 2am again and I am running the AI decision around in my head for the third night this week. Same loop. The investment you have not signed off, the team waiting on a direction you have not landed, the competitor who might already be three moves ahead. You reach for your phone, read one more article, and the loop tightens rather than loosens. By 3am you have not made the decision. You have only made yourself more tired for the meeting where the decision will fall to you.

Here is the thing worth knowing before you try to sleep: the loop is not waiting on a fact. If a missing number were the problem, you would have found it by now. What keeps you awake is uncertainty you have not yet metabolised, and the body files unprocessed uncertainty under "urgent" and runs it on a timer set for the small hours. The way through is not more information. It is a clearer state to meet the information you already hold.

Why is AI suddenly the thing keeping CEOs up at night?

Because the data caught up with the feeling. In the Conference Board's Q1 2026 Measure of CEO Confidence (the Conference Board is a long-running non-profit business research group whose quarterly survey tracks how chief executives read the economy), 60% of Fortune 500 CEOs named AI as the leading risk to their industry, up 7 percentage points on Q4 2025 and the first time AI has topped that list. In its C-Suite Outlook for 2026, the Conference Board found 38% of US CEOs identified AI as the leading societal, demographic or technological factor that could harm their business in 2026, ahead of political polarisation and trust in government. You are not catastrophising at 2am. You are replaying the same worry that now sits at the front of nearly every boardroom you respect.

And it lands in the body. The Calm Health executive survey found almost half of C-suite leaders, 48%, feel overwhelmed, only one in four say their mental battery is fully charged, and 41% report difficulty sleeping. The report makes the link plainly: leaders are struggling to switch off, making high-stakes calls while their reserves are drained, and the sleep-deprived ones are far more likely to feel overwhelmed. So the decision steals the sleep, and the lost sleep makes the next decision heavier. That is the loop you now inhabit.

Why does one more dashboard never quiet it?

Because the worry is internal, and you keep treating it as external. There is a quieter literature here worth your attention. Researchers describe decision fatigue (the way the quality of our choices degrades after a long run of deciding) as the impaired ability to decide and self-regulate after repeated acts of deciding; when those reserves run low, people lean harder on shortcuts, default, or avoid. The strong version of that mechanism is contested, so hold it loosely. But you already know the lived version. By the time the AI question arrives at 2am, it is the hundredth decision of a long day, arriving at your worst hour. No dashboard fixes the hour. It only gives the loop fresh material to chew.

Look at where the actual difficulty sits and the reframe gets easier. McKinsey's global survey found 88% of organisations now use AI in at least one business function, up from 78% a year earlier, yet adoption alone moves nothing: the redesign of workflows had the single biggest effect on whether a company saw bottom-line impact from generative AI, and a CEO's own oversight of AI governance was the factor most strongly correlated with that impact (EBIT, or earnings before interest and tax, is the standard measure of operating profit). Deloitte puts the same truth in starker terms. In its work on the AI value gap, it found companies still spend 93% of their technology funding on the technology itself and just 7% on training and upskilling the people who have to make it pay. The bottleneck is no longer the technology. It is the human and organisational layer, which is exactly the layer no amount of reading at 3am will resolve.

What actually settles the loop?

A clearer inner state, built on purpose. Coherence, the alignment between head and heart that lets you think well under pressure, is something you can develop, and there is defensible physiology beneath it. A systematic review of heart rate variability (HRV, the natural beat-to-beat variation in your pulse that reflects how well the nervous system is regulating itself) found that higher resting vagally-mediated HRV is associated with better cognitive performance, with executive functions showing the strongest link, consistent with the neurovisceral integration model (the theory that the vagus nerve and the prefrontal cortex work as one system to govern self-control). A later meta-analysis confirmed the same association, though the effect is small, so I name it as a direction to build toward rather than a magic switch. The principle holds either way: a regulated nervous system makes a cleaner decision than a depleted one. Steady the state and the same AI question stops presenting as an emergency.

So here is what I would do with the next 2am, in order:

1. Name it as uncertainty, not a fact gap. Write one sentence: "The thing I cannot yet hold is X." Naming it moves the worry from the body's urgent file to the mind's pending one.
2. Set the decision down until you are coherent. A choice made on a drained battery at 3am is the choice you will revisit tomorrow anyway. Give it your charged hour, not your worst one.
3. Steady the state first, then the strategy. A few minutes of slow breathing shifts the physiology that governs clear thinking. Decide from that calmer place.
4. Aim your attention at the people-and-process layer. That is where most of the value sits, and it is the part you actually own. Lead from where you intend the organisation to stand.

This is the deeper pattern underneath the loop. We have lived through three phases of how human output is won. Phase One, the Age of Effort: work hard, get a little more, linear growth. Phase Two, the Age of Scale: build once, sell to millions, exponential growth. Phase Three, the Age of Acceleration: output decoupled from human effort almost entirely, the phase AI now opens. Phase One was muscle. Phase Two was machine. Phase Three is mind, and the leaders who win this decade are the ones who upgrade themselves first. The AI decision will still be there in the morning. What changes is the state you bring to meet it.

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.