Anthropic CEO Dario Amodei worries that AI could eliminate half of entry-level white-collar jobs within five years.
There are many other potential labor risks, including the threat of massive U.S. tariffs on imports, layoffs aimed at redressing pandemic-era overemployment, and policies that empower employers at the expense of the workforce.
But Amodei seems to think policymakers are “bringing to the surface” the potential for massive job destruction from AI. This is a careful interpretation of the Trump administration’s no-questions-asked AI policy, which has rescinded Biden-era AI safety rules through executive order and praised AI infrastructure deals.
Amodei expressed this concern in an interview with Axios last week during Anthropic’s first developer conference, Code with Claude. He suggested the unemployment rate could reach 10-20% in the next 1-5 years.
“We have an obligation and an obligation as producers of this technology to be honest about what is going to happen,” Amodei told Axios. “I don’t think this is what people are interested in.”
Indeed, this concern appears in various reports, including the World Economic Forum’s Future of Jobs Report 2025. [PDF]. The report cites a poll of nearly 11,000 business executives that shows concerns about the impact of AI on employment are gaining some traction among business leaders.
Almost half of organizations expect to reorient their business models to new opportunities powered by AI
“Nearly half of organizations expect to reorient their business models towards new AI-driven opportunities (49%), while 47% plan to transition employees from roles disrupted by AI to other roles,” the employment report states. “Although most employers plan to hire new talent with AI-related skills, a significant proportion (41%) also expect to reduce their workforce as role-replicating AI capabilities expand.”
Changes are expected. “On average, workers can expect two-fifths (39 percent) of their existing skill sets to be transformed or rendered obsolete between 2025 and 2030,” the jobs report states.
As AI tools become more widespread, workers will need to learn new skills, which is always the case when technology reshapes society.
The introduction of AI does not necessarily mean that a wide range of jobs will disappear and new opportunities will not arise. As IBM CEO Arvind Krishna recently told the Wall Street Journal, “We’ve done a tremendous amount of work internally at IBM to leverage AI and automation in certain enterprise workflows, and our total employment numbers have actually increased because that frees up more investment to put into other areas.”
Regardless of the extent to which AI is creating new jobs at IBM, it’s important to note that the company’s current job postings indicate that those roles are taking place outside the United States, in low-wage regions such as India.
Academics also have concerns, but not to the extent expressed by Amodei.
In a recent paper titled “Artificial Intelligence and the Labor Market,” economists Menaka Hampol (Yale University), Dimitris Papanicolaou (Northwestern University), Lawrence D.W. Schmidt (Massachusetts Institute of Technology), and Brian Seegmiller (Northwestern University) found that AI will have a limited impact on labor as the labor market adapts.
“Overall, we find that AI has a modest impact on employment due to countervailing effects; high-exposure occupations are in relatively low demand compared to low-exposure occupations, but this results in increased productivity for firms and overall employment growth for all occupations,” the authors conclude.
said co-author Papanicolaou, a professor of finance at Northwestern University’s Kellogg School of Management. register It is difficult to predict the future and easy to predict what has already happened.
“In this paper, what we’re doing is looking at past episodes of what people called AI in 2010. AI is essentially just machine learning, big data, or essentially prediction, which is purely data-driven predictive algorithms,” he explained. “So I think that’s something completely different than what Anthropic is doing right now.”
Papanicolaou said the goal of the study is to distinguish between what it means for a job to be exposed to AI and what that means for the labor demands of that particular job.
“People tend to think that if their job is exposed to AI, it will replace my job and I will lose my job,” he explained. “So what we’re saying is, that’s not necessarily the case. There are some countervailing factors.”
Papanicolaou said the key is the extent to which AI can perform all the tasks involved in a particular job. If AI makes everything okay, the demand for labor could decrease. But it’s not such a bad thing if AI is better at some aspects of a human job but worse at other responsibilities.
“That could act as a countervailing factor, because it basically means that workers have time to focus on tasks that are not exposed,” he said. “In short, it’s not just the average exposure of your job that matters, but how it’s distributed across the different tasks you’re doing.”
As we reported last month, economists Anders Humlum and Emily Vestergaard, in a paper titled “Language models are big, labor market impacts are small,” found that AI chatbots will have little impact on 11 different occupations in Denmark in 2023 and 2024.
Asked about Amodei’s concerns, Humrum said: “Dario Amodei’s predictions are thought-provoking and certainly worth considering. After all, he is one of the people who has the best view of what technology lies ahead of us.”
“But we’re two and a half years into the widespread adoption of AI chatbots in the economy, and when we look at the data, these tools aren’t actually making a huge difference to employment or income across any profession.
Even the most innovative technologies, such as the steam engine, electricity, and computers, took decades to have a large-scale impact on the economy.
“Furthermore, looking at broader history, even the most innovative technologies, such as the steam engine, electricity, and computers, took decades to have a large-scale impact on the economy. Moreover, even when these technologies dramatically changed production, they did not cause mass unemployment in the long run.”
Why does Amodei make such a bold prediction when there is so much uncertainty and disagreement about the long-term impact of AI on jobs? Well, from a marketing perspective, it never hurts to portray your product as having the magical ability to change the world. And that may help him and the company secure a place in the regulatory arena. ®
