Last December, Anthropic did something that sounds a little weird but turned out to be brilliant: they turned Claude into an interviewer and let tens of thousands of users just talk.
No surveys. No multiple choice. Just a conversation with an AI about what they actually want from AI, what they’re afraid of, and how it’s already changing their lives.
80,508 people from 159 countries, speaking 70 languages, participated. That’s not just a lot of data—it’s the largest multilingual qualitative study on AI attitudes ever conducted. And the results are far more interesting than the usual tech punditry.
Hope and alarm live in the same person
The most striking finding? People aren’t neatly divided into “optimists” and “doomers.” Almost everyone carries both hope and fear, often about the same thing.
A lawyer from Israel put it perfectly: “I use AI to review contracts, save time… and at the same time I fear: am I losing my ability to read by myself? Thinking was the last frontier.”
That tension runs through the entire dataset. People want AI to handle drudgery, but they worry about losing their own skills. They want it to make them more productive, but they’re anxious about what happens when the machines get too good.
What people actually want
Anthropic used Claude to classify responses into categories. Here’s what people said they wanted most from AI:
Professional excellence (18.8%) was the top answer. People want AI to handle the boring, repetitive parts of their work so they can focus on the stuff that actually matters. A healthcare worker in the US described receiving 100-150 text messages per day from doctors and nurses. AI lifted the documentation burden, and suddenly they had more patience and more time to actually talk to families.
Personal transformation (13.7%) came next. This one surprised me a bit. People are using AI not just for work, but for emotional growth, therapy-adjacent conversations, and even learning how to be better humans. A respondent from Hungary said: “AI modeled emotional intelligence for me… I could use those behaviors with humans and become a better person.” That’s not what you expect from a tech survey.
Life management (13.5%) and time freedom (11.1%) rounded out the top categories. People want AI to handle the mental load of scheduling, reminders, and executive function. A manager from Denmark said: “If AI truly handled the mental load… it would give me back something priceless: undivided attention.”
The fears are real and specific
The concerns people raised weren’t abstract sci-fi scenarios. They were grounded in what’s already happening.
A technical support specialist from the US said: “I got laid off from my job in May because my company wanted to replace me with an AI system.” That’s not a hypothetical. That’s last year.
A software engineer from South Korea put the longer-term worry bluntly: “Humanity has never dealt with something smarter than itself. We need to reflect on how to prepare for the AI age.”
And then there’s the quieter fear that doesn’t make headlines: the slow erosion of skills. The lawyer who worries about losing the ability to read independently. The freelancer who’s been using AI to compensate for a chronic health condition but wonders what happens when the crutch becomes a dependency.
What’s missing from the public conversation
Most AI discourse is driven by people who don’t actually use the technology much. Pundits argue about AGI timelines. CEOs promise productivity gains. Regulators warn about existential risk.
What’s missing is the voice of the person who’s already living with AI every day and has started to develop an actual relationship with it. That’s what this study captures.
An entrepreneur from Nigeria said: “I live hand to mouth, zero savings. If I use AI smarter, it may help me craft solutions to that cycle. It still depends on me.”
That last part—”it still depends on me”—is worth sitting with. There’s no utopian fantasy here, no belief that AI will magically solve poverty. Just a practical hope that a tool might help, if used wisely.
The method matters
Anthropic used Claude itself to conduct the interviews and analyze the results. That’s not just a flex—it’s a genuine methodological innovation. Traditional qualitative research forces you to choose between depth and scale. You can interview 50 people in depth, or you can survey 50,000 people with shallow questions.
This approach bridges that gap. The AI interviewer adapts follow-up questions based on responses, so you get the richness of a real conversation at massive scale. Then Claude classifies the responses across multiple dimensions.
Is it perfect? No. The researchers acknowledge limitations: Claude’s classifiers might miss nuance, the sample is self-selected from Claude users, and there’s always the question of whether people talk differently to an AI interviewer than a human one. But for a first-of-its-kind study, the results are remarkably coherent.
What I take away from this
Two things stand out.
First, the dominant narrative about AI is too binary. The people who actually use it are living in the gray zone. They see both the promise and the peril, often in the same session. That’s not confusion—it’s wisdom.
Second, the most powerful use cases aren’t the ones that make headlines. They’re the quiet ones: a healthcare worker who can finally focus on patients instead of paperwork. A freelancer who’s been misdiagnosed for nine years and finally gets answers. A person in Hungary who learns emotional intelligence from a machine.
That’s not the AI future we’re sold in press releases. But it might be the one that actually matters.
Comments (0)
Login Log in to comment.
Be the first to comment!