Worried that AI might replace you? Check out this graph from Anthropic showing the jobs most at risk

• Anthropic economists say they have developed a new way to track how AI will upend the workforce.
• Their new measure shows that AI use hasn't come close to tapping the full power of large language models.
• Not surprisingly, they say the most exposed field is computer programmers.

Anthropic economists say that AI use is far from reaching its full potential to disrupt the labor market.

Using their new measure, they found the five most exposed occupations to be: Computer programmers, customer service representatives, data entry keyers, medical record specialists, and market research analysts and marketing specialists.

AI has yet to significantly affect the unemployment rate for workers in these highly exposed professions, economists Maxim Massenkoff and Peter McCrory wrote. The pair said there is "suggestive evidence" that the hiring of young workers in those fields has slowed.

Massenkoff and McCrory also wrote that there are a number of tasks and, in some cases, whole jobs that AI can't do, such as making legal arguments in a courtroom.

"Many tasks, of course, remain beyond AI's reach—from physical agricultural work like pruning trees and operating farm machinery to legal tasks like representing clients in court," the pair wrote.

The core of Massenkoff and McCrory's paper proposes a new way to measure AI displacement risk that combines real-world data on Claude usage with other factors, including tasks that are theoretically possible for AI.

Anthropic has been publishing real-world data on Claude usage for every state and Washington, DC, through their "Anthropic Economic Index."

By doing so, the pair said that they hope to pinpoint economic disruption more reliably in real time, making it easier to "help identify the most vulnerable jobs before displacement is visible."

"This approach won't capture every channel through which AI could reshape the labor market, but by laying this groundwork now, before meaningful effects have emerged, we hope future findings will more reliably identify economic disruption than post-hoc analyses," they wrote.

The measure, which they call Observed Exposure, shows just how far LLMs have to go to disrupt specific job tasks that AI could theoretically replace or augment.

"For instance, Claude currently covers just 33% of all tasks in the Computer & Math category," they wrote.

Dario Amodei has warned about the future of white-collar work

Anthropic CEO Dario Amodei has repeatedly sounded the alarm about AI job displacement. He has said that AI could replace up to half of all entry-level white-collar jobs in the next one to five years. Amodei has stuck by his views even as others in the industry, including OpenAI CEO Sam Altman, have questioned his outlook.

Massenkoff and McCrory's findings dovetail with a growing consensus that AI could eliminate most entry-level software engineering jobs. One of the biggest uses for Anthropic's Claude is coding.

Boris Cherny, creator of Claude Code, recently said he expects the title of software engineer to start to "go away" in 2026.

xAI CEO Elon Musk said last year that "anything that is physically moving atoms" will outlast AI disruption longer. The Anthropic economists found that the least exposed professions include cooks, motorcycle mechanics, lifeguards, bartenders, and dishwashers

It is worth noting that sweeping predictions of AI job disruption haven't always aged well.

Geoffrey Hinton, the so-called "Godfather of AI," said in 2016 that "people should stop training radiologists now" and that within five years AI would surpass humans in the field. A decade later, radiologists remain in demand. Hinton told The New York Times in 2025 that his prediction was too broad and that the timing was off, even as he was correct about the direction of AI progress.

AI disruption also won't affect everyone the same way, the Anthropic economists wrote.

Based on US Census Bureau data from the three months before ChatGPT's release, the economists found that "Workers in the most exposed professions are more likely to be older, female, more educated, and higher-paid."