Andrew Yang has been warning for years that automation would transform the U.S. workforce.
In the run-up to his bid for the 2020 Democratic presidential nomination, he predicted that self-driving cars alone could replace a million truck drivers, mostly men with limited education, a change that could “destabilize society” and even spark “riots on the streets,” he told The New York Times in 2018.
Seven years later, Yang says the crisis he feared is no longer hypothetical.
“Unfortunately, we’re aging very nicely,” he told CNN’s Michael Smerconish in a recent interview.
Job losses due to AI are no longer theoretical
AI and automation can already perform a significant portion of labor tasks in the United States and are reshaping employment, according to a recent analysis.
MIT’s Iceberg Index, released last week, found that current AI systems can already technically perform skills equivalent to 11.7% of the U.S. labor market (roughly $1.2 trillion in wages across financial, medical, and professional services).
The New York Times reported in October, citing internal Amazon strategy documents and interviews, that the company believes automation could help it avoid hiring more than 600,000 U.S. workers over the next few years, and that its robotics team has an ultimate goal of automating 75% of its work.
Salesforce, Walmart, HP, IBM, and Fiverr have all cited AI in recent layoffs or announced layoffs related to AI.
“44% of jobs in the United States are either repetitive manual tasks or repetitive cognitive tasks, so they could be targeted by AI or automation,” Yang said in an interview with CNN. “We’re seeing that play out now.”
Yang’s 44% estimate is broadly consistent with several major studies on automation.
In 2024, IMF analysis suggests that around 60% of jobs in developed countries will be affected by AI, with half benefiting from the technology and the rest being negatively impacted.
A McKinsey Global Institute (MGI) report released last month found that technology could theoretically automate more than half of American work hours.
Yang warns: Up to 40 million jobs could disappear
Yang pushed back against Roman Yampolsky, a University of Louisville computer science professor and AI safety researcher who predicted in September that there would be a 99% unemployment rate within five years.
“The situation will get worse. I certainly don’t think it will get worse 99% of the time,” Yang said.
Yang offered a rough estimate using a vulnerability benchmark of 44%. If the U.S. “mass-loaded” even half of those jobs over the next 10 years, the country could eliminate 30 million to 40 million jobs.
“It would be devastating,” he said. “It would be devastating for many communities.”
His fix: Guaranteed income paid by companies that win the AI boom.
The rapid pace of disruption caused by AI has revived Yang’s signature policy ideas. It’s a universal basic income that would give every American adult $1,000 a month without conditions.
He said the cash guarantee would help workers weather the shock of automation and maintain basic economic stability.
Funding for nationwide programs should be borne by the companies driving AI’s explosive growth, Yang said.
He cited Anthropic CEO Dario Amodei, who in May proposed to Axios the idea of a “token tax” for governments to impose on AI companies.
With big tech companies creating “hundreds of billions” of value, driven in part by data that the public has never intentionally provided, an AI tax or “computing tax” could collect “very large numbers very quickly,” Yang said.
Given that the U.S. generated about $85,000 in GDP per capita in 2024, according to the World Bank, a $12,000 annual free dividend “seems pretty modest and reasonable,” Yang said.
final warning
Mr Yang said cash transfers were not a panacea and people also needed purpose, structure and a sense of community.
But without financial stability, he said, millions of people risk being left behind and driven to radicalization.
“We can do so much more for the millions of Americans who will be evacuated,” Yang said.
