Fears about AI replacing white-collar jobs have been intensifying, especially after Anthropic CEO Dario Amodei recently warned that AI could eventually eliminate a large share of such roles. Now, new data released by the company provides a clearer look at which occupations may face the highest exposure as AI tools become more capable and widely used. Anthropic has published a list of the top 10 job roles that appear most exposed to AI automation, based on how frequently AI systems already perform tasks associated with those professions. The analysis comes as part of the company’s effort to track the labour market impact of generative AI systems such as Claude.
Anthropic reveals list of top 10 job roles most impacted by AI
At the top of the list are computer programmers, with the company estimating that about 74.5 per cent of their work tasks could be handled by AI tools. According to the data, the tasks most likely to be automated include writing, updating and maintaining software programs, activities that modern AI coding assistants are increasingly capable of performing.
Customer service jobs also feature prominently. Customer service representatives rank second on the list, with around 70.1 per cent exposure. AI systems are already being used to handle tasks such as answering customer queries, processing orders and responding to complaints through chatbots and automated support tools.
Another role with high exposure is data entry, where roughly 67.1 per cent of tasks could be automated. These tasks typically involve reading source documents and entering information into digital systems, a process that AI tools can now perform more quickly through automated data extraction.
Healthcare administration roles also appear on the list. Medical record specialists show an estimated 66.7 per cent exposure, with AI capable of compiling, summarising and coding patient data in medical record systems.
Beyond these roles, several business and analytical jobs also appear among the most affected. Market research analysts and marketing specialists show about 64.8 per cent exposure, as AI tools can help prepare reports, analyse datasets and convert complex findings into written summaries.
Sales roles are also included in the analysis. Sales representatives working in wholesale and manufacturing, excluding technical and scientific products, show 62.8 per cent exposure, with AI able to assist with customer outreach, product demonstrations and order management.
Financial sector roles also feature on the list. Financial and investment analysts have an estimated 57.2 per cent exposure, largely because AI systems are increasingly capable of analysing financial data and producing forecasts about businesses, industries or broader economic conditions.
Some technology jobs beyond programming also appear among the most exposed occupations. Software quality assurance analysts and testers show about 51.9 per cent exposure, with AI capable of detecting errors and suggesting performance improvements in software systems.
Meanwhile, information security analysts have an estimated 48.6 per cent exposure, as AI tools are increasingly used to conduct risk assessments and monitor security vulnerabilities.
Rounding out the top ten are computer user support specialists, with around 46.8 per cent exposure. AI chat systems and automated troubleshooting tools are already helping companies respond to common software and hardware problems raised by users.
The list is part of Anthropic’s latest attempt to measure how AI systems are interacting with real-world work tasks. Rather than simply estimating theoretical capabilities, the company says it looked at how often AI tools are actually performing tasks linked to specific occupations.
The findings arrive shortly after Amodei warned that rapid advances in artificial intelligence could eventually disrupt a large portion of white-collar employment. His comments added to an ongoing debate among economists and technology leaders about whether AI will mostly assist workers or replace them.
At the same time, Anthropic’s research also suggests that widespread job losses have not yet materialised. While some professions appear highly exposed in terms of task automation, the company has said earlier analyses show little clear evidence that workers in these occupations are currently facing significantly higher unemployment. Instead, the changes may first appear in more subtle ways, such as slower hiring for entry-level roles or companies relying more on automation to handle routine work.
– Ends
