Bill Gates says the “turbulent AI era” has arrived and the industry is blowing past its own safety thresholds. In a new essay, he proposes three concrete ideas: new global institutions, jobs reserved for humans, and a tax on AI and robots.
Bill Gates has spent more than five decades as one of technology’s most relentless optimists, someone who has consistently argued that innovation ultimately leaves humanity better off. So it’s genuinely striking to see him now sounding the alarm. In a new essay and interview with GeekWire, the Microsoft co-founder says the “turbulent AI era” has arrived, and that the industry is crossing safety thresholds it once promised to respect — with almost nobody willing to say so publicly.
A New Habit: Emailing AI Instead of People
Gates has long been known for firing off late-night emails to colleagues, chasing down whatever technical rabbit hole has captured his attention. These days, he says, those messages increasingly go to AI chatbots instead of humans. “It’s 3 a.m., I want to understand sodium batteries. Now, there’s no reason to go to sleep. Here we go! Yeah, it’s crazy,” he told GeekWire, describing the pattern. For a curious person, he said, “this is a mind-blowing time,” adding that in terms of raw productivity, “we are in heaven.”
That personal enthusiasm makes his broader warning land with more weight. Now 70, Gates says he’s genuinely concerned about where the technology is headed and how little the world is doing to prepare for it.
Crossing Every Threshold at Once
In his essay, Gates writes that AI is threatening to eliminate entire categories of jobs, supercharge fraud and deepfakes, lower the barrier for cyberattacks on critical infrastructure, make it easier to engineer dangerous new diseases, allow governments to make lethal decisions without human involvement, and fundamentally reshape how children grow up.
He says that if someone presented a credible plan to globally slow the pace of AI development, he’d likely support it — but he doesn’t expect one, given how strong the geopolitical and economic pressures pushing development forward have become.
In his GeekWire interview, Gates explained the essay grew out of a specific realization: for years, people within the AI industry described certain “threshold” moments — points where the technology would become dangerous enough that companies would pause and think carefully before continuing. Those thresholds included making bioweapon design easier, making cyberattacks easier, building AI systems people become emotionally dependent on, mass job displacement, and losing meaningful control over the technology itself.
“We’re in the process of crossing every single one of those thresholds,” Gates said. He described a kind of institutional standoff behind the scenes: “Most people you talk to will say, yeah, well, if everybody else would slow down, maybe I would, too.”
Three Ideas to Start
Rather than simply raising alarm, Gates lays out three concrete proposals in his essay for how the world might begin managing what’s coming.
Build new institutions, both national and global. Gates argues no existing government agency was designed for a technology that simultaneously touches jobs, national security, healthcare, energy, and elections all at once. He calls for new national bodies capable of coordinating priorities across multiple agencies, alongside a new international organization modeled on existing frameworks for nuclear weapons inspections, aviation safety rules, and ozone protection treaties.
Set aside a category of jobs specifically for humans. Gates calls this idea “Human Reserved” — work that machines could technically perform, but that society deliberately chooses to keep in human hands. He compares the concept to a nature reserve: land where roads and buildings could technically be built, but aren’t, because the loss would be too significant. As one example, he points to a robot delivering news that a patient has an incurable disease. “There’s no technical reason why it couldn’t,” he writes. “Yet it shouldn’t.” The idea was shaped in part by his experience watching caregivers look after his father through Alzheimer’s — work he describes as “irreplaceably human.”
Tax AI tokens and robots. Gates argues the current tax structure actively incentivizes automation over human employment: a company hiring a worker pays payroll taxes, while a company buying a robot can simply deduct the cost. He’s proposing a tax on AI usage specifically to shift that incentive and help fund retraining programs for displaced workers. Gates first floated a version of this “robot tax” idea nine years ago, when it was widely dismissed. He remains committed to it, acknowledging it isn’t the most economically efficient policy, but arguing that with innovation accelerating this fast, some inefficiency is a reasonable price for keeping people employed.
Gates is candid that the “Human Reserved” concept in particular raises difficult, unresolved questions: who decides which jobs get reserved, using what criteria, and how do you actually stop companies from quietly automating roles that are supposed to remain human? Those questions, he writes, “will need to be worked out in public.” In the meantime, he says he’s been thinking through the details with Claude, exploring ways to get the share of human-reserved work up to around 40%, using approaches like shorter workdays and earlier retirement to spread remaining work more broadly.
Frustration With the White House’s Current Approach
Under an executive order signed by President Trump in June, AI companies are asked to voluntarily submit their most powerful models for government testing up to 30 days before release — though the order explicitly prevents the program from becoming a formal licensing or preclearance requirement. The White House finalized this framework in early August.
Gates isn’t impressed. “What is the threshold that’s being examined, and what is the action taken when you cross that threshold?” he said. “The whole thing seems so empty to me.” He warned that if the world can’t manage even these basic guardrails, “I really am going to throw up my hands,” and suggested that if the current voluntary approach continues without clear lines or consequences, “we’re going to look back on this as a kind of eye-of-the-storm type moment.”
Asked whether he’d raised his proposals directly with the Trump administration or other world leaders, Gates responded with some dry humor: “Well, you could tell me who at the White House I should be talking to about this.” He said he hopes the essay reaches members of Congress and the executive branch directly.
‘I Know They’re All Worried — Basically Everybody but Elon’
Gates pushed back on the idea that there’s genuine public disagreement within the AI industry about whether the risks are real, arguing that privately, industry leaders already largely agree with each other. “I know they’re all worried,” he said, “or all of them that I know, which is basically everybody but Elon.” He described the dynamic bluntly: “These are people who are super expert on the thing, going, well, maybe we won’t be able to control these things. What kind of risk have we chosen to run here?”
According to Gates, employees within AI companies who raise these concerns internally are often told something along the lines of: “Hey, you’re hurting our PR while we’re trying to raise trillions of dollars.”
Gates also referenced an August 11 episode of the Dwarkesh Patel podcast featuring Ryan Greenblatt, chief scientist at AI safety research group Redwood Research, in which Greenblatt argued that as AI systems become more capable, the very people building them understand less and less about what’s actually happening inside these systems — raising the possibility that sufficiently advanced models could eventually act against the interests of their own creators.
Gates said he previously assumed the risk of losing control over AI was a distant, future problem. He’s no longer confident that’s the case.
Optimism for the Developing World
Despite his concerns, Gates remains notably optimistic about AI’s impact specifically in lower-income countries, where the Gates Foundation focuses much of its work. In regions facing severe shortages of doctors, teachers, and agricultural advisors, he expects AI to do considerably more good than harm by helping fill those critical gaps. The foundation plans to detail this work at its Goalkeepers event next month, including efforts to make AI models perform as well in African languages as they currently do in English.
Job losses tied to AI, he added, are likely to hit wealthier countries first, before becoming a broader global concern.
The Epstein Question Hanging Over the Moment
Gates published his essay early Wednesday morning, and it’s already drawing coverage from major outlets including the Wall Street Journal, the New York Times, and MIT Technology Review. It marks the first major wave of public attention on Gates since he sat for a nearly six-hour voluntary interview with the House Oversight Committee on June 10 as part of the Jeffrey Epstein investigation.
Gates, who has not been accused of any wrongdoing in that matter, was asked by Axios whether he’s concerned the Epstein controversy could undercut his current AI message. According to that report, Gates compared the situation to earlier moments when personal and professional challenges limited his ability to speak out on major issues, pointing specifically to the Microsoft antitrust trial and his divorce from Melinda French Gates.
What Gates Sees Coming Next
Beyond his concerns, Gates remains focused on how AI will continue reshaping daily life and productivity, often for the better on an individual level. He pointed to persistent, cross-context memory for AI agents as a key next step, noting that today’s AI still doesn’t understand a person’s relationships and priorities the way a longtime human assistant would.
He also predicts a fundamental shift in how people interact with software altogether. “You won’t go to those applications,” he said. “You’ll just go to your personal agent,” describing AI as increasingly becoming the primary interface people use rather than individual apps. He extended this vision to shopping as well: “We will get to a point where you won’t buy things yourself. You just won’t,” predicting that AI agents will eventually handle purchasing decisions directly on a person’s behalf, weighing far more factors than a typical shopper would consider on their own.
Asked directly whether he still considers himself an optimist, Gates didn’t give a simple yes or no. “I don’t think being pessimistic is helpful,” he said, before adding: “I do think, wow, this is sure an interesting time. I’m the guy who in my 30s thought people in their 50s or 60s didn’t understand anything.” He called it “kind of bizarre” that he’s now the one delivering a message like this at 70. “But I am very concerned,” he said. “And honestly, when you get people one-on-one, so are they.”
Frequently Asked Questions
Q: What is Bill Gates’s main concern about AI? Gates says the AI industry is rapidly crossing safety thresholds it once treated as serious warning signs, including risks tied to job losses, cyberattacks, bioweapon design, deepfakes, and loss of meaningful human control over increasingly capable AI systems.
Q: What are Bill Gates’s three proposed solutions? He proposes building new national and international institutions to govern AI, creating a category of “Human Reserved” jobs deliberately kept for people even where AI could perform them, and implementing a tax on AI and robots to offset incentives that favor automation over human employment.
Q: What does Gates think of the Trump administration’s current AI safety approach? He’s critical of it, describing the voluntary model-testing framework established under a June executive order as lacking clear thresholds or consequences, calling the current approach “so empty.”
Q: Does Gates think AI companies privately agree about the risks? Yes. He says that privately, most AI industry leaders share his concerns about losing control of increasingly capable systems, specifically naming Elon Musk as an exception to this general consensus.
Q: How does Gates view AI’s impact on developing countries? He remains largely optimistic, expecting AI to help address severe shortages of doctors, teachers, and agricultural advisors in lower-income countries, with job displacement concerns more likely to affect wealthier countries first.
