How People Are Actually Using AI, With Receipts

The AI conversation is loud. Depending on who you ask, it’s either about to replace everyone or already disappoint everyone. What’s been missing is something subtle and far more useful: evidence. Not opinions. Not think pieces. Just a look at how people are actually using AI in the real world.

That’s what the Anthropic Economic Index sets out to do. Instead of asking people what they think AI might become, it looks at usage patterns. What people ask for help with. Where AI shows up in real work. Across regions. Across tasks. Across contexts.

It doesn’t read like a breakthrough moment. And that’s kind of the point.

What This Index Is Actually Measuring

The Index looks at how AI systems are used in practice. Writing assistance. Summarization. Explanation. Coding help. Problem-solving. It’s based on observed behavior from anonymized Claude usage data, not surveys or self-reported optimism. That distinction matters, because behavior is harder to exaggerate.

If you’ve lived through a few tech cycles, this will feel familiar. New tools rarely arrive as revolutions. They arrive as helpers. Over time, those helpers quietly change how work gets done.

What People Are Mostly Using AI For

Across the board, AI is being used less as an autonomous worker and more as a supporting brain. People ask it to rewrite drafts, explain concepts, summarize long material, and debug code. It smooths rough edges. It fills in gaps. It speeds up thinking.

If you were expecting to see widespread job replacement in the data, you won’t. What shows up instead is augmentation. Human-in-the-loop everywhere.

North America: Productivity First

In North America, AI usage clusters heavily around knowledge work. Writing, research assistance, software development, documentation. AI fits neatly into desk-based workflows where speed, clarity, and iteration matter.

Here, AI often looks like a productivity layer. Something that helps people move faster through work they already understand. Less experimentation, more optimization.

Europe: Structure and Guardrails

European usage patterns show a more cautious relationship with AI. There’s greater emphasis on explanation, review, and controlled use. AI is often treated as something to be guided, checked, and constrained.

This lines up with Europe’s regulatory posture more broadly. AI is useful, but it’s expected to behave. Less improvisation. More structure.

Asia: Scale and Efficiency

Across parts of Asia, AI shows up strongly in operational contexts. Translation, workflow assistance, repetitive task support. The focus is less on ideation and more on throughput.

Here, AI is often about making large systems run more smoothly. It’s not flashy. It’s practical.

Africa and Other Emerging Regions: Access and Learning

In many African countries, as well as parts of Latin America and the Middle East, AI usage leans heavily toward learning and access. Explaining unfamiliar topics. Translating across languages. Filling educational gaps where formal resources are limited.

Here, AI isn’t primarily about optimization. It’s about opportunity. It acts as a bridge to knowledge, skills, and information that might otherwise be harder to reach.

That distinction matters. It shows how AI’s value depends less on the tool itself and more on the environment it lands in.

What Isn’t Showing Up

What the Index largely doesn’t show is widespread full automation. AI replacing entire roles outright remains rare in actual usage. People prompt. People review. People decide.

Change is happening, but it’s incremental. Quiet. The kind that doesn’t make headlines until years later.

Why the Receipts Matter

The value of the Anthropic Economic Index isn’t that it predicts the future. It’s that it grounds the present. It shows that AI adoption is uneven, contextual, and shaped by local realities.

It also reframes the conversation. Instead of asking whether AI will change work, the better question becomes who is already benefiting from it, and who still isn’t.

Receipts don’t settle every debate. But they do make it harder to argue from imagination alone.