Anthropic Reveals Which Jobs Are Most at Risk from AI
Anthropic built an early warning system to track AI's impact on jobs. Computer programmers, customer service reps, and analysts top the list.
Anthropic Reveals Which Jobs Are Most at Risk from AI
Anthropic — the company behind the Claude AI assistant — has built what it calls an “early warning system” to monitor which jobs are most at risk from AI automation. The research published this week reveals that many well-paying, white-collar professions are already in AI’s crosshairs.
What Anthropic Found
The study ranks jobs by how many of their tasks AI could potentially speed up or take over. Topping the list are computer programmers (75% task exposure), customer service representatives (70%), and data entry keyers (67%). Medical record specialists, market research analysts, financial analysts, and software testers all make the top ten.
Interestingly, the workers in these high-exposure roles tend to be older, female, more educated, and higher-paid — a profile that runs counter to the popular assumption that AI mainly threatens low-skill work.
At the other end of the spectrum, physically hands-on jobs like groundskeepers, cooks, motorcycle mechanics, lifeguards, and bartenders rank among the least exposed. These roles require human presence, physical coordination, and real-world judgment that AI still struggles with.
The Nuance: Exposure Isn’t Replacement
Anthropic is careful to point out that “exposure” doesn’t mean job loss. The research found “limited evidence that AI has affected employment to date.” A job being exposed means AI could assist with many of its tasks — not that the role disappears overnight.
That said, hiring trends are shifting. There’s early evidence that companies are hiring younger workers more slowly in roles with high AI exposure, suggesting the impact is beginning, even if it isn’t yet dramatic.
What This Means for You
If you work in one of the high-exposure fields — software, data analysis, customer support, finance, or marketing — the message isn’t to panic. It’s to adapt.
The workers most likely to thrive are those who learn to use AI tools effectively rather than compete against them. Tools like Claude, ChatGPT, and Gemini can handle repetitive tasks, draft documents, and analyse data — freeing you to focus on the judgment calls, client relationships, and creative work that AI can’t yet replicate.
Anthropic’s report is a signal worth taking seriously. Understanding where AI is heading in your field is the first step to staying ahead of it. You can read the full CBS News coverage of the study for more detail.
Want to keep learning?
Explore our guided learning paths or try building something with AI right now.
More from News
Adobe Firefly Custom Models — Train AI on Your Own Art Style
Adobe Firefly Custom Models — Train AI on Your Own Art Style
Adobe's Firefly Custom Models hit public beta on March 19, letting any creator train a personal AI on their own images. Here's how it works.
Visa Agentic Ready — AI Agents Can Now Buy Things for You
Visa Agentic Ready — AI Agents Can Now Buy Things for You
Visa launched Agentic Ready in Europe, letting AI agents make purchases on your behalf with 21 major banks already testing the system.
Tencent Is Building AI Agents Into WeChat for 1.4 Billion Users
Tencent Is Building AI Agents Into WeChat for 1.4 Billion Users
Tencent confirmed plans to embed AI agents in WeChat that can hail rides, shop, and book restaurants — bringing agentic AI to the world's largest super-app.
Enjoyed this article?
Subscribe for more AI insights delivered to your inbox every week.