What do you want from AI? Not what the tech companies want, and not what the headlines say. You. Anthropic, the company behind Claude, is asking that question right now, and anyone who uses Claude can answer it.

On September 29, Anthropic announced a new study where Claude interviews you about how you use AI and what you'd like from it. Here's how it works, what 81,000 people said the last time they did this, and one thing to do before you start.

What is Anthropic Interviewer?

It's Claude, set up to ask the questions instead of answer them. It asks about your experience using AI, what role you want it to play in your life and in the world, and what you want from the companies building it. Then it asks follow-up questions based on what you say, a bit like a real interview.

Anthropic's page says it takes about 15 minutes. And it says the goal is for researchers, policymakers and AI companies, Anthropic included, to learn from what people share and act on it.

What 81,000 people said last time

Last December, Anthropic ran the same kind of interview with 80,508 Claude users in 159 countries and 70 languages. Here are the headlines:

  • Most were positive. 67% leaned positive about AI overall, though most people held hopes and worries at the same time.
  • The top wish: do my job better. That was 18.8%. Next came personal growth (13.7%), keeping life organized (13.5%) and getting time back (11.1%).
  • The top worry: wrong answers. 26.7% worried most about AI being unreliable. Jobs and the economy came next at 22.3%, then losing control over decisions at 21.9%.
  • It's already helping. 81% said AI had moved them at least a little closer to what they were hoping for.

Here's what stands out to me. The top wishes aren't robots running the world. They're doing your job better, keeping life organized and getting time back. That's the same list of things most of us already use Claude for. And the number one worry, wrong answers, is the one you can manage yourself by checking what Claude tells you.

How to take part, and what to keep private

  1. Open the interview page. It's linked from Anthropic's post on X, and it opens inside Claude at a page titled "Your voice can help guide AI." Here's the link.
  2. Sign in to Claude if it asks.
  3. Read the short explanation, then click "Start here." Anthropic says participants are told how their answers will be used before they agree, so read that part.
  4. Answer like you'd answer a friend. Short and honest is fine.

One thing before you start: Anthropic says your answers help build a record of public responses that anyone can read. So leave out your full name, your employer, and anything private about your family or health. Talk about your experience, not your identity.

Is it worth 15 minutes?

It's your call, and it's optional. One interview won't change a product overnight. It's one input among many, and Anthropic hasn't said what it will change. But if you've ever wished somebody building these tools would ask what regular people want, this is somebody asking.

The takeaway

AI gets talked about a lot by people who build it. This is a chance for the people who use it for emails, paperwork and everyday work to be heard. If you do it, keep it personal in experience and private in details.

Want to see what everyday Claude help looks like, start to finish? That's what I do over on YouTube at @DoWithClaude: one simple task at a time, in plain English.