Elon University’s Imagining the Digital Future Center surveyed 1,505 American adults in May and published the results this week as a report called Human Purpose in Peril, written by Lee Rainie and Janna Anderson.
Two numbers from it sit next to each other and do not agree.
Sixty-five percent said they had used a large language model in the past week, up from 52 percent in early 2025. Sixty-four percent agreed that by replacing human beings in most tasks, AI will undermine our sense of purpose and meaning in life. Thirty-six percent took the opposing statement, that by freeing us from routine tasks AI will let us find greater purpose.
The rest of the findings run the same way. Sixty-nine percent said AI will weaken human-to-human bonds a lot or somewhat. Fifty-five percent are not confident their fellow humans will retain the ability to think critically and evaluate information, against fifteen percent who are. Fifty percent are not confident most people will use AI appropriately and effectively, again against fifteen. Forty-six percent are not confident people will maintain motivation and purpose as AI advances. Sixty-three percent said AI already heavily influences their daily routine or expect it will within five years.
So: two thirds use it, two thirds fear what it does to meaning, and these are overwhelmingly the same people.
The reflex is to call that hypocrisy. It is not. It is the most ordinary structure in consumer behaviour and coffee has been living inside it for twenty years.
Watch what people say about their high street. Ask anyone what they want and you will get the independent shop, the owner who knows their order, the room where the neighbourhood runs into itself. Ask them where they actually bought coffee this morning and a great many of them will name the chain, or the drive-through, or the machine in the office, and they will not be lying about either answer.
Both are true because they are answers to different questions. The purchase answers what is convenient at eight fifteen on a Tuesday when you are late. The preference answers what kind of place you want to live in. Individual convenience is decided one transaction at a time by one person. The character of a neighbourhood is decided by all the transactions together, and no individual transaction feels like a vote.
That is precisely the gap the Elon numbers are measuring. People are not confused about AI. They are making a rational individual decision about their own Tuesday and expressing an accurate collective worry about the sum of everyone’s Tuesdays, and there is no contradiction between those two positions at all.
Which is why the finding matters commercially rather than philosophically. When a population reports high adoption and high unease simultaneously, the unease is a market signal, not noise. It says a substantial share of people would pay something for the alternative if the alternative were actually available at the moment of decision. The independent cafe that survives is not the one that wins the argument about what a high street should be. It is the one that is open, on that corner, at eight fifteen.
Rainie’s framing of the underlying concern, that the worry lands on the very attributes people assumed would be our saving grace, is the sharpest part of the work.
Nobody in that survey has stopped using the thing. That is not the same as being happy about it.
