The Biggest Threat to Your AI Strategy Isn’t the Technology

A 26-year-old man stumbled into an emergency room convinced he had overdosed on his antidepressants. He was part of a clinical trial for a new drug and believed he’d consumed too many pills. His blood pressure cratered. His heart raced to 110 beats per minute. Doctors rushed to stabilize him, pumping nearly six liters of IV fluids into his system over the next hour.

Yet, he soon learned that he’d been assigned to the placebo group. Every pill he’d swallowed was chemically inert. Within fifteen minutes of learning the truth, his vitals stabilized.

This case appears in Nir Eyal’s new book, Beyond Belief: The Science-Backed Way to Stop Limiting Yourself and Achieve Breakthrough Results. Eyal argues that beliefs don’t just describe reality; they actively shape it. They influence our physiology, our motivation, and our capacity to change.

This isn’t a book about artificial intelligence. But it might give you an honest diagnosis of why your AI transformation is falling short of expectations—or why, despite the massive investment, it’s currently flatlining on the table.

Labels become prophecies

Eyal’s core insight is that beliefs don’t stay in your head. They calcify into labels. And once a label takes hold, it often inflicts more damage than the condition it was meant to describe.

Eyal points to a 2024 study by psychologist Nick Haslam and colleagues that found that when someone was tagged with the clinical label “Generalized Anxiety Disorder,” both the individual and those around them grew more pessimistic about recovery. But when the same condition was rebranded as a set of behaviors (“feels anxious,” “has trouble concentrating”), people were more optimistic.

Nothing about the underlying symptoms changed. What changed was the label. And that, in turn, reshaped what people believed was possible.