A report published last month by AI giant Anthropic made some crucial revelations, as it says workers in the most exposed professions are more likely to be older, female, more educated, and higher-paid.
The report was based on real-life enterprise usage of Anthropic’s popular Claude large-language model and it broke down a host of professions by their “observed exposure” and “theoretical exposure” to AI.
Essentially, it shows what share of the work Al systems can already do in a given occupation, and how much more they could theoretically take on.
Fortune says the report stoked worries about a white-collar recession. This is because a wide range of secure and well-paying white-collar occupations are now facing very high theoretical exposure.
These occupations include computer programming, market research, and financial management.
Here is a breakdown of the top ten occupations most exposed to AI automation, ranked by their Observed Exposure percentage and the specific tasks most likely to be automated.
Let’s break them down from the highest percentage to the least exposed jobs.
High-Exposure Occupations (65%–75%)
In this segment, we show the roles that face the highest level of task overlap with AI capabilities. The impacted jobs here are those in coding and direct communication.
At the top of the list are computer programmers (74.5%), who write, update, and maintain software programs that are facing automation.
Customer service representation (70.1%) is second, involving the automation of interacting with customers to provide information, take orders, and handle complaints.
They are followed by data entry keyers (67.1%), Medical Record Specialists (66.7%), and this includes compiling, abstracting and coding patient data.
Moderate-to-High Exposure Occupations (50%–65%)
Note that exposure in these fields centres on data synthesis, reporting, and specialised analysis, such as Market Research Analysts & Marketing Specialists (64.8%), which comes on top of the list.
Preparing reports, illustrating data graphically, and translating findings into text are now getting heavily automated.
Sales Representatives (Wholesale & Manufacturing) (62.8%) are next, with the automation of tasks like contacting customers to demonstrate products and solicit orders.
Then come Financial and Investment Analysts (57.2%) whose jobs of analysing financial information to forecast business, industry, or economic conditions are also edging towards AI adoption.
Software Quality Assurance Analysts and Testers (51.9%). They modify software to correct errors or improve performance.
Moderate Exposure Occupations (Under 50%)
While still in the top ten, these roles involve security and troubleshooting tasks that currently show slightly lower automation coverage.
Information Security Analysts (48.6%) and Computer User Support Specialists (46.8%).
The conclusion here is that knowledge work—specifically roles involving data processing, routine coding, and standardised communication—is at the forefront of the AI transition.
This aligns with the trend mentioned in the report: these “exposed” roles are often held by highly educated, higher-paid professionals whose work are now at risk.
