324
Audio & Video Production313
Software Development229
Automation & Workflow207
Writing & Content Creation190
Marketing & Growth179
AI Infrastructure & MLOps149
Design & Creative154
Photography & Imaging146
Data & Analytics115
Voice & Speech123
Education & Learning120
Sales & Outreach114
Customer Support112
Research & Analysis86
Anthropic says Claude data shows AI is not causing broad job losses yet, but hiring for some entry-level roles may be slowing in AI-heavy work.
In short: Anthropic says real-world Claude usage shows limited job displacement so far, but there are early signs that younger workers are finding it harder to enter some AI-exposed roles.
Anthropic’s February 2026 Economic Index report introduces “observed exposure.” It is a way to estimate how much AI is touching real work by combining what AI can do in theory with how people actually use Claude. Think of it like the difference between a kitchen that could cook 100 dishes and the smaller number it actually serves each day.
The report says most Claude use today covers only a small slice of the tasks AI could potentially handle. It also puts more weight on automated use through the API, which is a way for companies to connect Claude directly into software so tasks can run more automatically, like setting up a conveyor belt instead of handing tools to a worker.
Jobs with the most exposure include computer programmers, customer service representatives, and data entry keyers. Anthropic estimates computer and math roles have 35.8% task coverage in its data, and about 49% of jobs have at least 25% of their tasks showing Claude use. Even so, US survey data does not show a meaningful rise in unemployment for the most exposed jobs compared with less exposed jobs.
The report flags a possible “AI skills gap.” Since 2024, job finding rates into exposed occupations fell about 14% for ages 22 to 25, while older workers did not see the same drop. Anthropic also says coding use is shifting from the Claude website toward the API, which may favor experienced workers who can build automated workflows, and could widen inequality if entry-level hiring keeps slowing.
Source: TechCrunch AI