The paradox Keltner spent twenty years documenting

Dacher Keltner, a psychology professor at UC Berkeley, spent more than two decades studying what power does to people who acquire it. What he found across those studies was consistently uncomfortable: subjects under the influence of power acted as though they had suffered a traumatic brain injury — becoming more impulsive, less risk-aware, and less capable of seeing things from other people’s perspectives. He named the structural irony this creates the power paradox: once people have power, they lose some of the capacities they needed to gain it in the first place.

Displays of empathy and social intelligence are typically what cause others to trust someone enough to grant them authority. The paradox is that power then begins to erode precisely those qualities. Powerful people stop automatically mimicking others’ behaviours — tensing when others tense, laughing when others laugh — and it is exactly this behavioural mirroring that enables the automatic emotional transmission established in previous articles in this series. Without it, the powerful are progressively pushed toward an empathy deficit not through choice but through neurological consequence.

For entrepreneurs, the paradox operates across the growth curve. The founder who built the company on customer intuition, authentic human connection, and genuine attentiveness to others is, through the process of scaling and accumulating authority, being neurologically pushed away from the qualities that generated the insight the business was built on.

The neural mechanism Obhi found

Sukhvinder Obhi, a neuroscientist at McMaster University, studied the neural process underlying Keltner’s behavioural findings. Using transcranial magnetic stimulation — a technique that can temporarily disrupt specific neural processes — he measured the mirroring activity of powerful and less powerful participants and found that power measurably impairs the mirroring process. Power diminished the ability to read emotional expressions from photographs and to predict others’ reactions to specific words and actions.

The mirroring impairment is the mechanistic bridge between the power condition and the empathy deficit. The mirror neuron system is the neural substrate through which emotional contagion operates — the automatic process by which one person’s emotional state is transmitted to another’s nervous system through mimicry. When power impairs mirroring, it impairs this entire transmission pathway. The powerful person is not choosing not to empathise. The neural machinery that would enable automatic, involuntary empathy is being suppressed by the power condition itself.

The mirror neuron hypothesis remains contested in the scientific literature, and the direct causal link between mirroring impairment and reduced subjective empathic experience involves inference beyond what the TMS data alone can establish. Obhi’s findings are the leading mechanistic account; they are not yet definitive.

The E-test and what it reveals

Galinsky and colleagues’ 2006 study across four experiments produced one of the most cited demonstrations of power’s effect on perspective-taking. In the first experiment, participants primed with high power were significantly more likely than those primed with low power to draw the letter E on their forehead in a self-oriented direction — oriented toward themselves rather than readable to the person facing them. The high-power condition had automatically reduced their inclination to adopt another person’s visual perspective.

In subsequent experiments, high-power participants were less likely to account for the fact that other people did not share their privileged knowledge, anchoring too heavily on their own vantage point. They were also less accurate at reading emotional expressions — demonstrating power-induced impairment in the basic perceptual task that empathy requires. Numerous independent laboratories have since replicated the core finding: power reduces cognitive perspective-taking, compassion, advice acceptance, and accurate understanding of others.

The finding that most directly matters for entrepreneurs is the advice acceptance result. Power-induced disinhibition reduces the tendency to genuinely update on others’ perspectives, even when those perspectives have been explicitly sought. The founder who asks for feedback and then does not genuinely integrate it is not necessarily being arrogant. They are exhibiting the documented perspective-taking deficit that power produces — an anchoring on their own vantage point with insufficient adjustment toward the information they received.

The echo chamber that builds itself

The power paradox therefore has a direct strategic cost beyond the interpersonal one. It produces an echo chamber effect not because founders consciously surround themselves with agreement, though sometimes they do, but because the neural machinery for genuinely updating on others’ perspectives has been suppressed by the very success that produced the power. Customer research, investor feedback, team input, and market signals are all being processed through a perspective-taking system that power has systematically degraded. The misalignment between what the entrepreneur believes is clear and what the team actually understands is not dishonest — it is the E-test operating at organisational scale.

The structural antidote

The power paradox operates most severely in founder-led organisations where accountability structures are weak — no functioning board, no genuine 360-degree feedback, no external advisors with real authority to challenge decisions. The autonomy that makes entrepreneurship psychologically rewarding simultaneously removes the accountability architecture that would counteract the power-induced empathy deficit.

Research on the boundary conditions of the power paradox offers one constructive finding: when power activates a sense of responsibility for others rather than authority over them, the perspective-taking deficit is partially reversed. Founders who have internalised a genuine psychological commitment to their team, customers, and stakeholders — not as ethical performance but as a felt sense of accountability — appear to be structurally more protected against the empathy erosion that power normally produces. The distinction is not between powerful and powerless. It is between founders who experience themselves as accountable to the people they lead and those who experience themselves as sovereign over them.

Deliberately building accountability into governance — a board with genuine challenge authority, formal feedback mechanisms, external advisors whose role is explicitly to push back — is not just sound corporate structure. It is the architectural antidote to a documented neurological risk.

A book worth reading alongside this

The Power Paradox by Dacher Keltner is the primary source for the twenty-year research programme this article draws on, and the most directly relevant single text available on this subject. Keltner writes with unusual honesty about what his own decades of research revealed — that the people most likely to acquire power are those with the greatest social and empathic capacity, and that those same people are then systematically pushed away from those capacities by the very power they earned. The book covers the neurological findings, the mirroring suppression research, the empathy deficit studies, and the accountability antidote in a way that is both rigorously grounded and genuinely readable. For any entrepreneur who has noticed themselves becoming less curious about others, less willing to update on feedback, or less attuned to the emotional states of the people around them — and who wants to understand why that might be happening and what can actually be done about it — this is the most direct and most honest starting point available.

Have questions about this article?

If any part of this article raised questions you want to explore further, courbot.co is built for exactly that. It is courben.co’s AI assistant, designed around the psychology of entrepreneurship. Ask it anything from this article.

This article is for educational and informational purposes only. Sources: Keltner, D., Gruenfeld, D.H. & Anderson, C. (2003), Psychological Review, 110(2), 265–284. Obhi, S.S. et al. (2012), Frontiers in Human Neuroscience. Hogeveen, J., Inzlicht, M. & Obhi, S.S. (2014), Journal of Experimental Psychology: General. Galinsky, A.D. et al. (2006), Psychological Science, 17(12), 1068–1074.