Microsoft co-founder Bill Gates says artificial intelligence could eventually move society into a period of “true abundance,” but the transition may involve up to 20 years of severe disruption. The central issue for companies, workers and policymakers is not only how much productivity AI can deliver, but how employment, income and the meaning of work will be redistributed.
Speaking on a recent podcast, Gates said the next two decades could become “one of the most turbulent periods in human history.” He warned that AI could create clear winners and losers at the same time, placing sustained pressure on existing labor markets.
Call centers and accounting may be among the first to change
Gates said white-collar roles such as telemarketing, telephone customer service and accounting could shrink sharply because AI can perform many related tasks at lower cost, at greater speed and with higher accuracy. His assessment is not limited to office work. As humanoid robots advance, some blue-collar jobs could also be taken over by automated systems.
He said that unless people choose to slow or halt the process significantly, future robots could help build homes, grow food and even perform surgery. In that scenario, many goods and services currently produced with substantial human labor may no longer require large numbers of workers.
The economic consequences would extend beyond corporate cost-cutting and productivity gains. If production requires progressively less human labor, wages, vocational training, the tax base and consumer purchasing power could all face adjustment. The speed at which existing jobs disappear compared with the pace at which new roles emerge will be a key measure of AI’s impact on the economy.
Work provides more than a paycheck
Gates also stressed that work matters for reasons beyond income. It can give people a sense of purpose, establish a daily routine, create relationships with colleagues and provide satisfaction through solving problems or making products.
He noted that even if AI eventually becomes better at activities such as art, sport, cooking and caregiving, societies may still choose to keep people involved in those fields. For Gates, the question is not simply whether machines can perform a task, but whether people want efficiency to become the sole standard for deciding how work is done.
Gates also envisioned that some people might adopt a lifestyle resembling that of the Amish, using AI only within limited boundaries in order to preserve mutual assistance and room for original ideas. That choice would not necessarily mean rejecting technology altogether; it could instead represent an attempt to draw a line between greater efficiency and continued human participation.
Shorter working hours emerge as one buffer
In response to the employment pressure AI could create, Gates has suggested that earlier retirement and limits on working hours might help preserve employment to some extent. Amazon founder Jeff Bezos and JPMorgan Chief Executive Jamie Dimon have also discussed shorter working weeks, arguing that they could provide a more gradual adjustment than mass unemployment.
Other technology executives have proposed different approaches. Elon Musk, the founder of Tesla and SpaceX, has said that AI could drive production costs sharply lower, potentially turning work into a choice. He has also suggested that governments could distribute funds to the public as employment declines. Under that scenario, people might have access to more goods and services, while traditional saving and retirement planning could become less central.
Dario Amodei, chief executive of Anthropic, has said AI could improve living standards for everyone, although the scale of the short-term disruption could be unprecedented. OpenAI Chief Executive Sam Altman said in May that AI had eliminated fewer white-collar jobs than he previously expected and that he no longer believed an “employment apocalypse” was approaching.
Unemployment has not risen in lockstep
Although large companies including Amazon and Meta have announced layoffs, concerns about AI replacing corporate roles have intensified without a clear corresponding rise in the national unemployment rate during the current phase of rapid AI development. Employment data so far do not establish that large-scale job replacement is already taking place across the economy, nor do they rule out effects emerging gradually over a longer period.
Gates’s view is therefore better understood as a warning about the future structure of the labor market than as a description of current unemployment data. As companies continue deploying AI and robotics, key questions will include which occupations decline first, whether productivity gains translate into broader income growth, and whether shorter working hours or earlier retirement can genuinely cushion the labor-market transition.