Everyone at the table agreed the product needed AI features. I was not sure anyone could explain why.
This was a meeting with a client, early 2025. The brief was straightforward: a dashboard for internal reporting. No unstructured data, no natural language queries, no use case that remotely required machine learning. But AI was in the brief because AI was in every brief that quarter.
I asked why. The room went quiet. Then someone said, "Everyone else is doing it."
In 1951, a psychologist named Solomon Asch ran an experiment that should be taught in every business school. He sat a group of people around a table and showed them three lines of obviously different lengths. Each person was asked, in turn, which line was longest. The answer was plain.
But every person at the table except one was an actor. And the actors, on cue, all pointed to the wrong line.
The real participant, the one who could see the correct answer clearly, pointed to the wrong line too. Seventy-five percent of participants gave at least one incorrect answer. When asked afterwards why, two explanations emerged. First: "I did not want to be the one who was wrong." Second: "Everyone else agreed, so I thought I must be the one seeing it incorrectly."
Asch summarised the result simply: when you take the truth away from someone's eyes, they will seek safety in the majority.
The technology industry runs on a version of this experiment, played out daily. A framework becomes popular and suddenly every new project uses it, regardless of whether it fits. A company succeeds with a particular model and every competitor copies it, regardless of whether their context is the same. A term enters the vocabulary and it appears in every pitch deck within months, regardless of whether anyone in the room can define it precisely.
The penalty for going along is hidden. You build something that works but is overcomplicated. You adopt a tool that is wrong for the job but defensible because everyone else is using it. You ship a feature nobody needed because the market "expected" it.
The penalty for dissenting is visible. You have to justify yourself. You have to explain why you are not doing what everyone else is doing. The burden of proof falls on the person who disagrees, never on the consensus.
I have been in rooms where I was the youngest person by a decade and the only one questioning the premise. That is not a boast. It is a description of what it feels like to dissent when the social cost is high: uncomfortable, lonely, and occasionally wrong.
But the times I went along, the times I pointed to the short line because everyone else did, those are the decisions I regret. Not because dissent is always correct. It is not. But because agreement without examination is not agreement at all. It is compliance.
The Turkish novelist Halit Ziya Uşaklıgil wrote, over a century ago: "People are strange. The moment they sense they are doing something different, they will find a reason to silence their own conscience."
He wrote that in 1897. Fifty-four years before Asch ran his experiment.
Some patterns do not need laboratories to be understood.