AI: Writers, Developers, and Designers in Imminent Danger

Digital jobs, such as writers, developers, and web designers, are at the forefront of occupations threatened by artificial intelligence. This is revealed by the American AI Jobs Risk Index, a groundbreaking study published by Tufts University in Boston. This index ranks 784 occupations according to their actual vulnerability to AI, based on data from the O*NET database, the Anthropic Economic Index, and research from Microsoft Research.

The study focuses on the American labor market and does not merely measure theoretical exposure to AI. It projects concrete job losses and quantifies them in terms of lost income. Although the results are centered on the United States, the occupations identified as the most vulnerable are similar to those held by digital professionals in France and Europe.

The 20 Occupations Most Threatened by AI Job Loss

The index goes beyond measuring exposure to AI; it projects the actual vulnerability of each occupation to job eliminations. Among the 20 most threatened occupations are writers and authors, with a risk of losing 57.4% of positions within the next two to five years. Computer developers and web and digital interface designers closely follow, with rates of 55.2% and 54.6% respectively. Editors, at 54.4%, and occupations in mathematical sciences, at 47.6%, are also among the most at risk.

Other professions, such as web developers (46.2%), database architects (46%), and operations research analysts (45.1%), are also under threat. Atmospheric and space scientists (44.2%), sociologists (43%), and technical writers (42.4%) complete this concerning picture. Statisticians (42.1%), political scientists (40.3%), and database administrators (39%) are not far behind.

Public relations specialists (37.3%), social science researchers (37.3%), and data scientists (37.2%) are also at risk. Personal financial advisors (37.1%), software QA testers (36.4%), and research assistants in social sciences (36.1%) are included in this list. Just behind them are market research analysts and marketing specialists (35.5%), information analysts, reporters, and journalists (35.4%), as well as interpreters and translators (34.2%).

The Most and Least Exposed Occupations to AI

Conversely, hundreds of occupations show a projected loss rate of 0%, such as masons, masseurs, roofers, stretcher-bearers, and fast-food cooks. Researchers summarize this situation by stating that "the jobs that AI cannot touch are largely those that the economy has always undervalued."

By sector, information services lead with 18.3% of jobs threatened, followed by finance and insurance (16.5%), professional, scientific, and technical services (15.6%), and business management (14.1%). Across the United States, the study projects 9.3 million jobs at risk in its median scenario, with a range from 2.7 to 19.5 million depending on adoption assumptions. The economic impact is estimated at $757 billion in annual revenue, equivalent to the economy of Belgium.

"The Promise of AI Productivity is a Displacement Pipeline"

The most striking finding of the study is encapsulated in a paradox, summarized by researchers in a formula: "The promise of AI productivity is a displacement pipeline." This means that the jobs where AI increases productivity the most are also those that show the highest job losses. The correlation is clear: for every additional percentage point of automation, the index projects a 0.75 point loss in employment.

Thus, when AI significantly enhances the efficiency of an employee, organizations can produce as much with fewer people, initially reducing hiring for junior positions. This dynamic reverses historical patterns of automation. It is no longer low-skilled manual jobs that are at the forefront, but cognitive-intensive occupations such as programming, writing, financial analysis, and design.

In terms of absolute revenue, the heaviest losses do not hit the same occupations as those ranked by percentage. It is software developers, management analysts, and marketing specialists who bear the brunt of the impact, due to their high salaries and the volume of workers involved.

Signs Already Visible in France

These American projections resonate with data from France. As we reported recently, an analysis by Insee shows that employment among those under 30 is already declining in the IT and information services sectors in France: a decrease of 7.4% year-on-year in the fourth quarter of 2025, while activity in these sectors continues to grow. The mechanism identified by Insee is the same as that described by Tufts, with an adjustment that first involves a slowdown in hiring for junior positions rather than layoffs.

The study enters an increasingly dense field of research. Anthropic introduced its own metric of actual exposure, the observed exposure, in March, while Microsoft Research published a similar ranking based on usage data from Copilot. The American AI Jobs Risk Index stands out by moving from exposure to vulnerability. While previous studies measured what AI could theoretically touch, Tufts projects what it is likely to concretely eliminate. The authors clarify that they deliberately excluded AI-related job creation from this version of the index due to a lack of sufficiently robust data.