dev.toJuly 9, 2026
Feature

35% of Workers Say AI Will Do Most of Their Job Within a Year. Anthropic's Own Data Proves Them Right.

Anthropic just published the most detailed look yet at how AI is reshaping work. The June 2026...

Anthropic just published the most detailed look yet at how AI is reshaping work. The June 2026 Anthropic Economic Index report, based on millions of Claude conversations and a survey of 9,700 users, reveals a workforce that knows the ground is shifting under its feet.

Over one third of respondents expect AI to handle most or nearly all of their work tasks within 12 months. The people who delegate the most to AI are the most optimistic about their careers. And the youngest workers are the most afraid of losing their jobs.

These are not predictions from pundits. This is data from the company that builds one of the world's most capable AI models, drawn from real usage patterns and direct survey responses. ([Anthropic Economic Index](https://www.anthropic.com/economic-index))

---

What the Economic Index Actually Measures

Anthropic launched the Economic Index in early 2026 to track how people use Claude across every US state and hundreds of occupations. The June 2026 report, titled "Cadences," is the fourth installment and the most ambitious. It introduces three new data sources: hourly usage sampling, an output classifier that labels what each conversation produces, and the Economic Index Survey. ([Anthropic, June 2026 Report](https://www.anthropic.com/research/economic-index-june-2026-report))

The dataset is public. Anthropic publishes it on HuggingFace, and the full methodology is in an [open appendix](https://cdn.sanity.io/files/4zrzovbb/website/03ed1410f74a65ae4cc2a27120d0875e1e569535.pdf). This is not a black-box study.

---

Finding 1: Your Workweek Is Etched Into AI Usage

Claude usage mirrors the workweek with striking precision. On weekdays, roughly 35% of conversations are personal. On weekends, that jumps to nearly 50%. Business correspondence, marketing copy, and slide decks give way to emotional support, medical questions, and investment advice.

The hourly data is even more revealing. People ask for news at 7 a.m. Business correspondence peaks at 10 to 11 a.m. Recipe requests spike to 2.3 times the average at 6 p.m. Sleep advice peaks around 5 a.m. Tax-related conversations surged to 8 times the average on April 14, the day before the US filing deadline, then dropped sharply on April 16.

On weekends, when people do turn to Claude for work, the tasks skew toward higher-wage occupations. Tasks related to jobs in the bottom two wage quartiles, like telemarketing and clerical work, shrink. This pattern holds even when you remove computer and mathematical occupations entirely. People in higher-paying jobs work more outside traditional hours, and they bring AI with them.

---

Finding 2: Higher-Wage Work Costs More Compute

Conversations mapped to higher-wage occupations consume more tokens. A lot more. Marketing managers earn roughly twice what editors earn ($80 vs. $37 per hour), and their Claude conversations consume approximately 2.5 times as many tokens.

This is not just because higher-wage work is more complicated. It i

Read full article on dev.to