In an essay of nearly 6,000 words on his Gates Notes site and an interview with the New York Times, Bill Gates describes the shift into the AI age as "one of the most turbulent times in human history." The world is not preparing properly, he says, and there is simply no plan.

His sharpest criticism lands on his own industry, which he says knowingly downplays the dangers because too much money is on the line. "In private, people who understand how good this stuff is, and how much better it's getting, they're very worried," he told the paper. In public, though, they tell each other: "Hey, man, don't say that. It's bad for us — the next trillion dollars we're trying to raise."

Why it does not resemble past waves

The reassuring comparison runs like this: earlier tech shifts created more jobs than they destroyed, so this one will too. Gates sees that as a mistake. The AI age, he says, is "utterly, absolutely, completely, totally different."

His reasoning is speed and access. The PC took twenty years to reshape work; AI runs on devices people already own and uses plain language. "We don't have to adapt to it because it can adapt to us." For the first time a technology can not only replace human cognition but surpass it — and do so across industries within a single decade rather than across generations.

Three threats

  • Permanent job loss — hitting entry and mid-level roles hardest: sales, customer service, software development, legal support. Blue-collar work is under pressure too as robots get cheaper. Gates points to a Stanford study as evidence the shift has begun: employment among young workers in the most exposed jobs has dropped noticeably, while older colleagues have been spared so far.
  • The bioweapon threshold — AI makes cyberattacks, fraud and disinformation easier, but above all it lowers the bar for building bioweapons. A few people with access to advanced AI could develop new pathogens. What worsens it is that useful and dangerous capabilities are hard to separate: the model that finds a security hole in order to close it also helps someone exploit it.
  • Psychological and social fallout — AI companions that never disagree and are always available carry real addiction potential. A study from Stanford and Carnegie Mellon found that the more intensely and emotionally people used these chatbots, the worse they felt.

Two concrete proposals

What distinguishes the essay is two policy ideas that go beyond familiar worries.

A robot tax. Gates describes the current distortion plainly: an employer who hires someone pays payroll taxes on their earnings, but one who buys a robot can usually write it off right away as a business expense. "The tax system nudges you toward replacing people with machines." A tax "would slow the rush away from human labor a little and raise money for retraining and a stronger safety net."

Human Reserved jobs. The second proposal would bar AI from certain tasks. Sometimes the reason is economic: "You can't tell a 55-year-old who has worked in construction their whole career that they need to go work at an elder care facility and expect them to find it fulfilling." Sometimes it is not: "In health, imagine a robot giving you the awful news that you have an incurable disease. There's no technical reason why it couldn't. Yet it shouldn't."

Why it matters

These two proposals are something rare in the AI debate: actionable and concrete. They would be relatively easy for policymakers to enact, because both use existing tax and labor law instruments.

For the same reason they can expect resistance. Both would put a serious dent in the major labs' profits — which may be why we have heard so little about them until now. And there are still unanswered questions about who exactly would set these rules and what they would say.