AI's Uneven Impact: Which Jobs Are at Risk and Which Are Safe?

A new study by U.S.-based AI firm Anthropic reveals how artificial intelligence is reshaping the job market. The research highlights a growing divide between AI's theoretical capabilities and its actual use in fields like IT, finance, and administration. It also shows which professions are most at risk—and which remain untouched by automation.

Anthropic's report introduces a fresh way to measure AI's impact: observed exposure. Unlike earlier studies relying on theory, this metric blends real-world data with existing models. It tracks how often AI handles tasks in knowledge-heavy roles, such as programming (75% exposure), customer service (70%), and legal documentation.

The findings point to a clear trend. Every 10% increase in AI-managed tasks trims a profession's expected job growth by 0.6 percentage points. White-collar jobs—software developers, financial analysts, and customer support staff—face the highest exposure. Surprisingly, workers in these sectors tend to be better educated, earn more, and are more likely to be women. Yet not all jobs are equally vulnerable. AI struggles in roles demanding physical presence or fine motor skills, like chefs, mechanics, or bartenders. Despite the risks, the study found no widespread layoffs in AI-prone fields as of early 2025. However, new graduates in these areas now encounter fewer job openings. Anthropic's approach also helps economists and policymakers tell long-term shifts from short-term market swings. The methodology allows for regular updates, ensuring the data stays current as AI adoption grows.

The study underscores a gap between AI's potential and its real-world use. While automation reshapes white-collar work, physical and skill-based jobs remain largely unaffected. For now, the biggest impact falls on hiring trends—not mass redundancies—particularly for those entering the workforce.