Are You Losing Your Job to AI? Anthropic Report Flags Programmers And Finance Workers As 'Most Vulnerable' Jobs

Anthropic has published research warning that computer programmers, customer service workers and financial analysts face the highest displacement risk from AI, but the data tells a more complicated story than the headlines suggest.

The paper, authored by economists Maxim Massenkoff and Peter McCrory and published on 5 March 2026, introduces a new framework for measuring AI's real-world impact on the labour market, one that deliberately separates theoretical capability from what AI is actually doing in professional settings today.

The findings offer some reassurance alongside several sharp warnings, and they arrive at a moment when anxiety about AI-driven job losses has moved well beyond the pages of academic journals and into boardrooms, parliament buildings and household conversations.

A New Way of Measuring the Risk

The starting point for the research is a problem that has long plagued attempts to forecast AI's economic impact: past predictions have repeatedly overstated disruption. Anthropic's economists point to a well-known 2009 study that identified roughly a quarter of US jobs as vulnerable to offshoring, a finding that, a decade later, had simply not materialised in employment data. The lesson they draw is that measuring what AI could theoretically do is not the same as measuring what it is actually doing.

To bridge that gap, the authors created a metric they call 'observed exposure.' The measure pulls from three sources: the O*NET database, which catalogues the tasks associated with around 800 US occupations; real-world Claude usage data gathered through the Anthropic Economic Index; and task-level exposure scores from a 2023 academic study by Eloundou et al., which assessed whether an AI model could theoretically double the speed at which a given task is completed.

Crucially, the new measure weights automated use, where AI is replacing a worker's output, more heavily than augmentative use, where a human is simply being assisted. A job where AI helps a worker do more is treated differently from one where AI has started doing the job outright.

Computer programmers sit at the top of the exposure ranking, with 75% of their tasks now covered by observed AI use. Customer service representatives follow, driven largely by the rapid rise of AI-powered chat interfaces deployed directly by companies via API. Data entry workers, whose core task of transcribing and processing documents has long been ripe for automation, are 67% covered. Financial analysts also rank among the most exposed, given how much of their work; data retrieval, modelling, summarisation, maps directly onto what large language models do best.

The Gap Between Theory and Reality

Despite those headline numbers, the research lands with a significant caveat built in. AI is far from reaching its theoretical capability: actual coverage remains a fraction of what is technically feasible. Even in the most exposed occupational category; Computer and Mathematical roles, Claude currently covers just 33% of all tasks, against a theoretical ceiling of 94%.

Office and administrative roles sit at a theoretical 90%, but observed coverage lags far behind. Legal constraints, software requirements, human verification steps and simple inertia are all slowing the pace of adoption, even in sectors where the technology could, in principle, take over tomorrow.

That gap matters enormously when it comes to unemployment. The paper's authors are explicit: they find no systematic increase in unemployment for workers in the most AI-exposed occupations since late 2022, when ChatGPT's release began transforming public awareness of the technology.

Using data from the US Current Population Survey, the economists tracked unemployment trends for workers in the top quartile of observed exposure against those with zero exposure. The gap between the two groups has remained statistically insignificant. In plain terms, highly exposed workers are not yet losing their jobs at a measurably higher rate than anyone else.

The research does, however, find one early warning signal that is harder to dismiss. Hiring of younger workers, specifically those aged 22 to 25, into highly exposed occupations has slowed. The job-finding rate for that age group entering high-exposure roles has dropped by roughly half a percentage point since 2022, representing a 14% fall compared to the pre-ChatGPT baseline.

The same decline does not appear among workers aged over 25. This echoes a parallel finding from a separate 2025 study by Brynjolfsson, Chandar, and Chen using payroll data from ADP, which found a 6–16% fall in employment among 22-to-25-year-olds in exposed occupations, attributed primarily to a slowdown in hiring rather than an increase in redundancies.

Who Is Actually at Risk — and Who Is Not

The demographic profile of the most exposed workers is, on the surface, counterintuitive. The most at-risk group earns 47% more on average than workers with no AI exposure. They are more likely to be female, more likely to hold a graduate degree, and more likely to be Asian or white.

People with graduate qualifications make up just 4.5% of the unexposed group but 17.4% of the most exposed group, an almost fourfold difference. This runs directly counter to the common assumption that AI threatens low-paid, low-skilled work first. The jobs currently in the crosshairs are well-compensated knowledge roles: the programmers, analysts and customer service managers who sit in open-plan offices and work through screens all day.

At the opposite end of the spectrum, 30% of US workers have zero coverage under the observed exposure measure, because their tasks appear too infrequently in AI usage data to register. That group includes cooks, motorcycle mechanics, lifeguards, bartenders and dishwashers — workers whose jobs depend on physical presence, dexterity and real-time human judgement. AI can do many things; it cannot yet bus tables or resuscitate a drowning child.

The paper also cross-references its exposure measure against the US Bureau of Labor Statistics' official employment projections for 2024 to 2034, published in 2025. The relationship is modest but consistent: for every 10 percentage point increase in observed AI coverage, the BLS's projected employment growth for that occupation drops by 0.6 percentage points. The direction is clear, even if the magnitude is, for now, small.

The question is no longer whether AI will reshape knowledge work, but how fast the uncovered gap between theoretical capability and actual deployment will close — and whether workers, companies and governments will be ready when it does.