Claude Users Talk About AI Like It’s Personal

Anthropic just published a large qualitative study based on 80,508 conversations with Claude users from 159 countries and 70 languages, looking at how people talk about and use AI. It’s one of the bigger datasets we’ve seen that focuses on user sentiment, not just performance metrics.

For context, most AI coverage focuses on what models can do, but this one looks at how people feel while using them. Different angle, same tech.

The patterns are familiar but still a bit surprising when you see them at scale. People use Claude for practical things like writing, coding, and research, but also for more personal situations. Talking through problems, asking for advice, even framing it like a second opinion. Hmm.

There’s a mix of curiosity and caution. Some treat it like a tool they control, others sound more like they’re collaborating with it. The language changes depending on the situation. Task mode versus conversation mode.

The results also highlight concerns that come up repeatedly. Trust, accuracy, overreliance. People know it’s not perfect, but they still lean on it more than expected. That shows up across different languages and regions, not just one user group.

So?

This is what early adoption actually looks like. Not just people using AI, but figuring out where it fits in their thinking.

The tech hasn’t changed overnight, but the way people relate to it is changing. That part is harder to measure, but probably matters more.

You'll find a link below for the full post.