Anthropic CEO Warns Half of White-Collar Jobs at Risk Over Next Five Years

The ability of AI to displace humans at various tasks is accelerating quickly, Anthropic CEO Dario Amodei warned at the Axios AI+ DC Summit on Wednesday. Speaking to government officials and tech leaders, Amodei reiterated his controversial prediction that up to half of white-collar jobs could be wiped out by AI over the next five years.

"As with most things, when an exponential is moving very quickly, you can't be sure. I think it is likely enough to happen that we felt there was a need to warn the world about it and to speak honestly."

— Dario Amodei, CEO of Anthropic

The warning comes at a critical moment as Amodei and others argue that the speed of AI displacement could require immediate government intervention to support displaced workers. Anthropic co-founder Jack Clark, who oversees the company's policy operation, emphasized that "you need some kind of policy response at the scale of disruption we expect in the next five years."

Unemployment Could Spike to 10-20%

In previous statements to Axios, Amodei predicted that AI could eliminate half of all entry-level white-collar jobs and spike unemployment to between 10% and 20% within five years. If realized, this would represent the highest unemployment rate since the Great Depression, fundamentally reshaping the American economy.

Amodei's warning encompasses a "white-collar bloodbath" that could affect mass elimination of jobs across technology, finance, law, consulting and other white-collar professions, with entry-level positions particularly vulnerable. The disruption represents a fundamental shift from historical patterns where technology typically automated blue-collar manufacturing jobs.

Entry-Level Career Pipeline Under Threat

The disruption of entry-level positions creates a particularly dangerous scenario. If new workers find it hard to land these roles, they may not develop their own workplace skills in a timely manner, disrupting the job market pipeline over the long term and affecting companies looking to hire skilled, experienced people in the future.

"Most lawmakers are unaware that this is about to happen. It sounds crazy, and people just don't believe it."

— Dario Amodei on government preparedness

Anthropic research reveals a critical transition happening in AI usage patterns. Currently, the company tracks that about 60% of people use AI models for augmentation—helping people do their jobs—while 40% use them for automation—actually doing the job. However, the automation percentage is growing rapidly.

Unprecedented Speed of Change

"It's going to happen in a small amount of time—as little as a couple of years or less," Amodei warns. This timeline is far more compressed than previous technological disruptions, leaving little time for workers, companies, or governments to adapt. The exponential nature of AI improvement means that capabilities can advance dramatically in short periods.

Agentic AI: The Game Changer

Hundreds of technology companies are developing agentic AI systems—AI agents that can perform human work "instantly, indefinitely and exponentially cheaper." These agents, powered by large language models like Claude, represent a fundamental shift from tools that assist humans to systems that can replace them entirely.

Amodei emphasizes that government intervention will be necessary to support displaced workers as AI rapidly transforms the job market. The Anthropic CEO advocates for transparency from both AI companies and governments about the workforce implications of advancing AI technology.

Anthropic's Transparency Initiative

Amodei has created the Anthropic Economic Index to provide real-world data on Claude usage across occupations, along with an Economic Advisory Council to help stoke public debate. He hopes other companies will share similar insights, giving policymakers a more comprehensive picture of AI's workforce impact.

"We, as the producers of this technology, have a duty and an obligation to be honest about what is coming."

— Dario Amodei on industry responsibility

Amodei acknowledges the apparent contradiction in his position—building increasingly powerful AI systems while simultaneously warning about their disruptive potential. This "very strange set of dynamics" reflects the complex reality facing AI leaders who see both the tremendous promise and peril of their technology.

Claude Writing Its Own Future

At the same summit, Amodei revealed that Claude is increasingly writing the code to build future versions of itself. "Claude is playing this very active role in designing the next Claude," he noted, explaining that the "vast majority" of future Claude code is being written by the AI system itself. This recursive self-improvement capability could accelerate AI development even further.

The Window for Action Is Closing

The urgency in Amodei's warnings reflects the compressed timeline for adaptation. Unlike previous technological shifts that unfolded over decades, AI capabilities are advancing exponentially, potentially leaving society little time to adjust. The choice, according to Amodei, is between proactive preparation now or reactive crisis management later.

As Amodei poses the critical question: "Well, what if they're right?" The AI leader's stark warnings may sound alarmist, but the consequences of ignoring them could be far worse than the discomfort of confronting them now. The challenge for policymakers, business leaders, and workers is to take these warnings seriously while there's still time to shape a more equitable AI-powered future.