A latest report from AI startup Anthropic reveals that enterprises are overwhelmingly using AI to automate work rather than engage in human-machine collaboration, intensifying risks that AI could disrupt the job market.

According to the research report released by Anthropic on Monday, more than three-quarters (77%) of enterprise usage of its Claude AI software involves automation patterns, often including "complete task delegation." This finding is based on analysis of Anthropic's application programming interface (API) traffic, which is primarily used by developers and enterprise users.

As one of OpenAI's main competitors, Anthropic is a leading company selling AI tools to enterprises, aiming to help accelerate tasks such as software development, research, and writing. However, this technology has also raised concerns about potential mass layoffs and employee displacement, a risk that Anthropic also emphasized in its report.

Regarding this latest data, Sarah Heck, Anthropic's head of external affairs, stated:

"We don't know if this confirms the company CEO's previous pessimistic predictions about the job market, but this data suggests that some new developments are occurring."

**Primary Use Cases: Administrative and Coding**

The report shows a surge in automation usage rates, but the exact driving forces behind this remain unclear. Peter McCrory, Anthropic's head of economics, indicated that it's currently unclear whether this growth is due to "new model capabilities expanding the scope of tasks that can be automated" or "people becoming more adapted to large language models and more willing to delegate certain tasks to Claude."

McCrory believes that determining exactly which factor is driving this shift "is an important area for future research."

The report further clarifies the specific areas where enterprises are automating tasks. Overall, Anthropic found that enterprises primarily use Claude to handle administrative tasks and programming work. Among these, coding has consistently been a key focus area for Anthropic and the entire AI industry.

AI developers including Anthropic and OpenAI have released more sophisticated AI tools capable of writing and debugging code on behalf of users. These tools directly target high-skill positions such as software development, and their enhanced automation capabilities further highlight technology's potential impact on white-collar jobs.

Notably, Anthropic's leadership has previously issued stern warnings about AI's disruptive potential. Company co-founder and CEO Dario Amodei has predicted that AI could eliminate 50% of entry-level positions and publicly stated that people should stop "sugar-coating" future challenges.

This newly released report provides data support for these predictions to some extent, quantifying the actual behavioral patterns of enterprise users who tend to use AI for task replacement rather than assistance. Although the company officially states uncertainty about whether this means the worst-case scenario will eventually arrive, this report undoubtedly provides key reference material for assessing AI's direct impact on the labor market.