AI Is Reshaping Jobs—but Not Replacing Them, New Study Finds

A new study by U.S.-based AI firm Anthropic has examined how artificial intelligence is reshaping the job market. Released in early 2026, the research tracks real-world adoption rather than relying solely on theoretical forecasts. It reveals that AI's impact varies widely across professions, with some sectors adapting more smoothly than expected.

One key finding is that AI tools like Claude Cowork are being used alongside workers rather than replacing them outright. Analysts at Bernstein noted on March 7, 2026, that the system integrates into existing software workflows as an 'intelligence layer', reducing fears of widespread job displacement.

Anthropic's study introduces a new measure called observed exposure, blending theoretical models with actual data on AI usage. This approach allows for regular updates as AI adoption grows. The findings show that every 10% increase in AI-handled tasks reduces a profession's expected job growth by 0.6 percentage points.

So far, AI's strongest effects are in 'white-collar' roles, particularly among software developers, customer support staff, and financial analysts. Workers in these fields tend to be highly educated, earn above-average wages, and are more likely to be women. Yet, despite concerns, no systemic job losses have been recorded in AI-vulnerable sectors as of early 2025. However, hiring rates for recent graduates in these areas have fallen by 14% since ChatGPT's launch. The study also highlights a stark divide in AI's capabilities. While it excels in data-driven tasks, it remains ineffective in jobs requiring physical presence or fine motor skills, such as chefs, mechanics, or bartenders. Bernstein's March 2026 report further confirmed that tools like Claude Cowork are being embedded into platforms like Salesforce and Workday as supportive aids rather than replacements. Anthropic's methodology aims to help policymakers separate long-term structural changes from short-term market shifts. The goal is to develop targeted support for workers in transitioning fields.

The research suggests AI's role in the workforce is evolving more collaboratively than initially feared. While hiring trends have softened in exposed sectors, outright job losses have not yet materialised. The study provides a framework for monitoring AI's expanding influence and preparing workforce adaptations as needed.