You're not late on AI

I did not feel triumphant. I felt a small, slightly embarrassing relief — then irritated that a headline was what it took.

I have been told I am late. Late to the chatbot. Late to the new button in my email. Late to the television that started offering to talk. Late is a polite word for behind. Behind is a polite word for the thing I have been pretending not to mind.

The headline did not make me feel caught up. It made me feel counted. Being counted by a poll is a thin kind of company. I sat with it anyway.

I went looking for the number underneath the sentence. Pew Research asked 3,488 U.S. adults, in the last week of June 2026, how the increased use of AI in daily life made them feel. 52% said more concerned than excited. 9% said more excited than concerned.

Here is the part that shifted it. That 52% is not a new mood that arrived this summer. It was 52% in 2023, 51% in 2024, 50% in 2025, 52% again. Pew has been asking since 2021, when it was 37%. The industry had four years of a majority-shaped unease and still wrote the story as if time would convert us. Time did not.

Sarah Perez, writing it up for TechCrunch in August, named the part I had been living without a citation. The features arrived in the inbox and on the television whether anyone asked. Brian Chesky, who runs Airbnb, said the backlash is real, and that they have not shipped products "regular people" can love. His example of a thing we would love was a doctor on demand. Dario Amodei, who runs Anthropic, called the sour mood a crisis of trust, and said the accurate criticism is that the companies have not yet delivered on the big promises.

I did not need a CEO to tell me the television started talking. I needed someone to stop calling the pause a skill gap.

The age split is the detail I keep turning over. Adults 65 and older: 59% more concerned than excited, the highest of the four bands Pew published. Adults under 30: 55%, a majority for the first time. The people who were supposed to do the converting are not doing it. The people who were supposed to be the holdouts were never a special case. They were just easier to scold.

I thought I was going to land somewhere cleaner than this. Something like: see, I was right. I don't have that.

I still open the chatbot. Concern and use can sit in the same week without one of them being a lie. Pew's younger adults are the heaviest users and the newly majority-concerned. That is not hypocrisy. That is what it looks like when a tool is handy and the bargain is still ugly.

The other wrinkle: adults 50 to 64 came in at 47% more concerned — the lowest of the four bands, not the highest. Even the story I wanted, the one where "our people" have been saying this all along, is messier than a slogan. The late/behind story needs someone to play the slow one. I have been volunteering for that part without noticing.

Chesky can call it a narrative problem. A narrative problem is a problem you fix with better sentences. I have heard better sentences. They still arrive inside a product I did not request, next to a data centre someone else has to live beside, under a promise that keeps moving while the thing in my hands is a summary of a webpage I could have read.

I am not late. I was early to a feeling they measured four years running and called a messaging problem. I will keep using the useful parts. I will keep not pretending the rest won me. Those are not contradictions.

I don't know where you are with all of this. I suspect you have been called late for a feeling you had on time. I'd be curious what you did with that.

Kevin note: If a product added an AI feature you did not ask for, look for the setting that turns it off. You do not owe the feature a trial.

A few terms came up in this one — chatbot, data centre. Both in the glossary if either is new to you.

If someone asks: "I read something that said regular people still aren't sold on AI, and that isn't a personal failing. The polls have been majority-concerned for years. Using a tool and not joining the applause can be the same week."