AI Can Do Most White-Collar Work But Isn't Yet — And When It Does, Lawyers, Analysts, and Developers Face the Greatest Risk
Summary
A new Anthropic study warns that AI can already handle up to 94% of tasks in fields like law, finance, and software development, but adoption lags far behind capability — and when it catches up, a 'Great Recession for white-collar workers' could follow, with early signs already emerging in declining hiring rates for AI-exposed roles.
Key Points
- Anthropic publishes a new study revealing that AI is capable of performing the vast majority of tasks in fields like business, finance, law, and computer science, but actual adoption remains a small fraction of what is technically possible.
- The workers most at risk when AI adoption catches up to its capabilities are older, highly educated, and well-paid professionals — such as lawyers, financial analysts, and software developers — with AI theoretically able to handle 94% of tasks in computer and math roles, yet currently covering only 33% in practice.
- While mass layoffs directly tied to AI have not yet materialized, researchers warn a 'Great Recession for white-collar workers' is possible, and early signs include a 14% drop in hiring rates for AI-exposed occupations among young workers in the post-ChatGPT era.