The Dichotomy of Control in the Age of AI

58% of workers report anxiety about AI replacing their roles, and yet the same studies reveal something more instructive: 73% of those same workers express frustration not about AI itself, but about having no voice in how it is implemented around them.

Read that again slowly. The numbers do not describe a fear of machines. They describe the oldest wound in working life—the experience of being acted upon rather than acting.

Marcus Aurelius wrote, in the second century, a line that belongs in every modern workplace: Some things are in our control and others not. He did not originate the thought—that was Epictetus, his philosophical forefather, a man who had been a slave and understood powerlessness not as theory but as daily reality. What Marcus did was live the principle from the opposite position: as emperor, surrounded by decisions being made in his name, about things he could not always influence. His Meditations are, among other things, a sustained practice in identifying what was genuinely his to govern and releasing his grip on what was not.

This is not resignation. Epictetus was explicit on this point. The dichotomy of control is not a philosophy of passivity—it is a philosophy of precise action. You stop exhausting yourself on things outside your jurisdiction so that you can act with full force on the things that are.

Now bring that lens to the current moment in organizational life.

The Anxiety Is Mislabelled

When workers say they are anxious about AI, we tend to interpret that as technological anxiety—fear of the algorithm, fear of automation, fear of being made redundant by a system that does not know their name. And that fear is real and deserves honest attention.

But the data suggests something more specific is happening. The frustration about lack of agency in AI implementation decisions points to a different source of suffering entirely. It is not primarily what AI will do. It is that decisions are being made about people's working lives without those people's participation.

This is Stoic territory. The pain of powerlessness is one of the most consistent sources of human distress across the philosophical traditions, and the Stoics mapped it with unusual precision. What makes powerlessness particularly corrosive is not merely that you cannot act—it is that you cannot see clearly what you could act on, because anxiety has collapsed the distinction between what is yours to govern and what is not.

In conversations with people navigating organizational AI rollouts, we hear the same formulation repeatedly: I don't know what's going to happen. That uncertainty is real. But when we probe further, the more precise statement often emerges: I don't know what I'm supposed to do about it. Those are different problems. The first is about prediction. The second is about agency.

The 14-Month Gap

We observe that the average gap between recognising a problem and taking meaningful action is 14 months. Fourteen months of living with something that troubles you, named but unaddressed. And 67% of users who describe feeling stuck report that the stuckness predates their awareness of it by six months or more—meaning the paralysis set in long before the person consciously registered it.

This is precisely what unexamined anxiety does. It does not announce itself cleanly. It accumulates in the background, shaping behaviour and narrowing options, until the person finds themselves in a position that feels both sudden and inevitable.

AI anxiety in the workplace follows this pattern faithfully. The technology arrives. The rollout is announced. People feel the ground shift and, lacking a clear framework for what is theirs to govern and what is not, do what anxious systems do: they either freeze or they protest the wrong thing. Neither relieves the actual source of discomfort.

What Marcus Would Observe

Marcus Aurelius governed an empire that included constant uncertainty—military, economic, political, philosophical. His Meditations were not written for publication. They were written as daily practice, a discipline of returning his attention to the distinction between what required his direct action and what required his acceptance. He was not seeking tranquility in the sense of comfort. He was seeking clarity, which is a different and more useful thing.

Applied to workplace AI anxiety management, the Stoic framework does not say: do not worry about AI. It says: identify, with precision, which aspects of your situation are genuinely within your governance.

You may not control whether your organisation adopts an AI system. You likely cannot reverse a boardroom decision already made. These are, in Epictetan terms, not up to you—and exhausting yourself trying to undo them is a waste of the capacity you actually possess.

What is up to you: how you build knowledge of the tools now, before the moment of forced adoption. How you position yourself as someone who engages rather than retreats. How you articulate, through whatever channels are available, your perspective on implementation. How you develop skills that travel regardless of what specific tools the organisation selects. How you establish your own criteria for what constitutes meaningful work, so that you are not purely at the mercy of an externally defined role.

These are not small things. They are the things that actually determine outcomes at the individual level.

The Examined Work Life

Epictetus taught that the unexamined response—the automatic reaction of fear or avoidance—is the primary source of unnecessary suffering. Not the event itself, but the unexamined relationship to it.

Workplace AI anxiety, at its root, is often not managed because it is not examined. People carry it as a generalised dread rather than a specific, bounded problem with specific, bounded responses. The Stoic discipline is to sit with the anxiety long enough to ask: what precisely am I afraid of, and which of those specific fears point to something I can act on?

Users who complete a concrete action within 48 hours of identifying a problem are 3.2 times more likely to sustain engagement with that problem over the following week. The principle is ancient and the data confirms it: clarity without action is not yet clarity. It is only when the examination produces movement that the anxiety begins to resolve.

Marcus wrote to himself, not for applause. The Meditations are a record of a man doing the work of self-governance daily, imperfectly, persistently. That practice—not the conclusions, but the practice—is what the current moment in working life requires.