Why we ask AI before asking people
Originally posted on LinkedIn, June 24, 2026.
I told my wife, “Ask the chat.” I wasn’t joking.
Do you have a library membership? Of course not. We moved to the web, then to asking people on social networks. Now another quiet but steady shift is changing our preferences, almost without us noticing: we would rather consult a machine than a person.
A few years ago, we asked our neighbors to recommend a good electrician. Today, we ask on Facebook. There was no dramatic turning point. At some point, the web simply gave us a faster, more accurate answer, and that was enough to change the habit. The same thing is happening now with AI.
I no longer go to Fiverr to find a developer for a small project. I have asked my product manager why she came to me, a person, with a question that an agent could answer in a minute and more accurately. During an argument with my wife, heaven forbid, I have already said, “Ask the chat.” These things are not happening because I am a true believer in technology. They are happening because it works.
I am an extreme example because I work at the leading edge of AI and spend my days deep inside it, so I encounter this shift more often. But the rest of the world is moving in the same direction. Even our parents are starting to consult the chat instead of Google or Facebook.
Take medicine. Roey Tzezana described visiting a specialist whom the chat itself had recommended after reviewing his test results and symptoms. The doctor agreed with the chat’s diagnoses. He noticed a consistent rise in a particular parameter that Tzezana’s family doctor had missed for five years in a row, just as the chat had. He recommended the same test, with the same warnings. His rule was clear: don’t go to a doctor without talking to AI first, and don’t trust AI without talking to a doctor.
Less than two years ago, the default was to ask a person and use a tool only when there was no other choice. Now that default is starting to reverse. We ask the machine first, then turn to a person if needed. It is not happening all at once or at the same pace in every field, but the direction is clear. Anyone who works with AI every day already feels it.
