AI companies often talk about the future. Anthropic decided to ask people living in it.
Anthropic conducted one of the largest qualitative studies ever, using its AI system to interview over 80,000 people across 159 countries and 70 languages in just one week.
That scale alone is unusual.
But the method is what stands out.
Instead of surveys, the AI held real conversations, adapting follow-up questions based on each person’s responses, creating something closer to human interviews than traditional data collection.
And what people said is not simple.
It is contradictory.
On one hand, many users described AI as genuinely helpful. One respondent said, “Claude put the historical pieces together, leading to my proper diagnosis after being misdiagnosed for over 9 years.”
Another shared a more grounded reality: “If I use AI smarter, it may help me… It still depends on me.”
That captures the tone of the entire study.
AI is useful.
But it is not fully trusted.
Because alongside the benefits came real fears.
Some participants spoke about job loss directly. One said, “I got laid off… because my company wanted to replace me with an AI system.”
Others worried about something less visible.
Loss of thinking ability.
Dependence.
Control.
Anthropic found that hope and fear are not divided between different groups.
They exist in the same person.
Someone can rely on AI to save time, and at the same time worry they are losing their ability to think independently.
That tension is the real story.
Because it suggests that AI adoption will not be driven by capability alone.
It will be driven by trust.
People are not just asking what AI can do.
They are asking what it might take away.
There is also a broader shift here.
This study shows that AI is not just being used for productivity.
Many people want it to help them live better, not just work faster, whether that means solving problems, learning new skills, or navigating personal challenges.
That expands AI’s role significantly.
From tool.
To support system.
But it also raises the stakes.
Because the more influence AI has in people’s lives, the more important its reliability, behavior, and alignment become.
Anthropic itself acknowledges this tension.
The usefulness is real.
The risks are real.
And both are growing at the same time.
So what this research really shows is not just how people use AI.
It shows how they feel about it.
And that feeling is not excitement or fear alone.
It is both.
So the real question is not whether AI will become more powerful.
It is whether it can become trustworthy enough for people to rely on it without constantly questioning what it might cost them in return.
