This is a memo by Bill Gates, published on GatesNotes on August 26, 2026, arguing that the transition to the AI era will be "one of the most turbulent times in human history" and that "right now, we are not preparing adequately for that transition." It is a position paper rather than a research finding, and Gates presents it as the first of a series. It carries three proposals, which he describes as a start rather than a plan.
The memo's own title is "The turbulent AI era is here. The choices we make now are critical.", with the standfirst "We need a plan to ensure that the good outweighs the bad." Coverage frequently used the URL slug, "a turbulent AI era and critical choices to make," which is not the memo's headline.
The argument
Gates's framing claim is a disjunction: "AI will either be the greatest equalizer ever invented, or the worst source of injustice." He states that he sees no evidence "that leaders, experts, and communities are confronting the challenges adequately," and that "there is no plan to ease the entry into the AI era."
Two arguments support the claim that this transition differs from earlier ones. First, on reliability: the visible errors that lead commentators to discount AI — the strawberry-letter-counting and Sudoku failures — are being fixed quickly "as researchers create models that can check their own work and improve themselves." Second, on adoption speed: the personal computer took twenty years to change how people worked because software had to be written, prices had to fall, and users had to learn the tools, whereas AI "runs on the devices we already have, and it uses natural language. We don't have to adapt to it because it can adapt to us." He contrasts the agriculture-to-office-work shift, which "proceeded over several generations and created new jobs where human cognition was required," with a technology that "can substitute for human cognition," and puts the timescale at "a decade rather than a few generations."
Gates discloses a potential bias in the text: that he has benefited from the technology industry and retains financial ties to it, that he works with Microsoft and other AI companies as chairman of the Gates Foundation, and that profits from his investments go to the foundation. "Of course, readers will have to decide for themselves whether this clouds my view."
He also states that he "would likely support" a credible plan for slowing AI advances globally, while not expecting one: "The geopolitical and economic incentives are pushing too hard to go full speed ahead."
Three risks
Jobs disappearing permanently. Gates cites roughly 25 percent US unemployment in 1933 as a reference point, arguing the AI effect "may not reach this level" but "will not go away with an economic cycle" because the jobs most at risk are entry- and mid-level while new jobs require years of training. He cites the Stanford Digital Economy Lab's "Canaries in the Coal Mine" finding that employment fell significantly among young workers in occupations vulnerable to replacement after generative-AI adoption, but not among older colleagues. Named early-affected occupations: sales and customer support, software engineering, and paralegal work, extending to loan assessment, data analysis, and patient triage. He notes software engineering may generate offsetting demand as costs fall. On blue-collar work, he expects "smart" robots to compete with people on physical tasks in construction and hospitality "by the end of the decade," and says much of the advanced dexterous-robotics work is being done in other countries, primarily China. The competitive mechanism he describes is a price spiral: one firm adopts, lowers prices, and competitors must follow or be displaced by start-ups. He identifies the decisive threshold as the point at which AI produces "nearly error-free work," after which "it will be able to function on its own without a human checking in on it, and companies will have every economic incentive to let it."
Empowerment of bad actors, and eventual loss of control. Gates argues AI lowers both the information and the execution barrier for harm, so that "even criminals with very limited skills will be able to target victims at every scale." On cyber, he states that "the smartest cybersecurity experts I know are scared about the next few years, because the attackers are getting powerful new capabilities faster than the defenders can fix all the weaknesses," and identifies the structural difficulty as inseparability: "the same AI model that can find a flaw in software so a company can fix it can also help a criminal exploit it." He makes the parallel argument for biology. He then notes that the same tools concentrate power where it already exists, naming autonomous weapons and cheaper opinion manipulation. The section closes on control: "AI systems themselves already occasionally act in ways their designers didn't intend … as the models become more powerful, they could begin to act against our interests and we could lose control."
Children's development and human relationships. Gates argues AI companions "don't push you outside your comfort zone," are always available, and "never get mad at you," giving them "the potential to become highly addictive." He cites a Stanford and Carnegie Mellon study of more than 1,100 AI-companion users finding that those with smaller social networks were most likely to turn to a chatbot, and that heavier and more emotionally personal use correlated with feeling worse. He quotes Jonathan Haidt's The Anxious Generation on children exposed to small risks growing into adults who handle larger ones, and adds: "An AI companion designed to never upset you is a big, protected greenhouse." On education he cites a preliminary survey associating heavier AI use with less critical thinking, with a stronger effect among younger people, and argues this is "the worst possible time for humans to lose their critical thinking skills." He names Australia, the United Kingdom, and Norway as adopting online protections for children and states that China "has gone the furthest," with rules that "restrict AI companion apps broadly, bar designs that foster emotional dependence, and ban virtual relatives and romantic partners for minors."
Three proposals
A new system for managing the transition, which Gates calls "the highest priority" and "a monumental task." His argument for new institutions is that existing ones each see only part of the problem: "A labor department may understand workforce disruption but not security risk. A business regulator may understand market concentration but not AI's effects on children and teenagers." His scale comparison is the post-9/11 reorganization of the US government, "its biggest reorganization since World War II for the purpose of improving just one function," where AI "will require much, much more" across national security, employment, education, taxation, energy, elections, air and water, public health, finance, law enforcement, transportation, public lands, and IT. At national level he wants bodies that set priorities across agencies, so that "an AI-enabled attack" does not succeed "because no one thought it was their job to stop it." Internationally he proposes an organization drawing on three named models — the nuclear-weapons inspections regime, international aviation regulation, and the ozone-layer agreements — needing "elements of all three and more," and states that "some cooperation between the U.S. and China will be required." He proposes that countries hosting leading developers and controlling critical supply-chain segments "should begin meeting now to set up shared norms, before competitive pressure makes it harder for them to cooperate."
Human Reserved. Gates proposes a domain of work set aside for people even where machines could do it, a term he says he has "started calling" this and offers as "an example of the kinds of ideas we'll need to consider." The analogy is to nature reserves: "places where we could put buildings and roads, but we choose not to because the loss would be too great." He grounds it in the paid caregivers who looked after his father in the later stages of Alzheimer's disease before his death in 2020 — "Something in the care they gave my dad was irreplaceably human. No robot could or should have done it." The two examples in the memo are elder care and delivering a terminal diagnosis: "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." He gives both an economic rationale (displacement of workers who cannot easily retrain — "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") and a non-economic one, expects the boundary to evolve and to vary by country, with Japan's shrinking workforce cited as a case that "may welcome a caregiving robot," and expects education and mental health care to be mixed, "with a human in charge who's using the technology to extend what they can do." He states the open questions himself rather than presenting the idea as settled: who decides what is reserved, on what criteria, how companies are kept from using robots anyway, and what happens to international trade when one country permits robot production and another does not.
Rebalancing labor and capital taxation. Gates proposes taxing AI tokens and robots, on the argument that displaced or under-employed workers pay less income tax exactly when safety-net demand peaks. His stated mechanism is that the current system is already tilted: "if you're an employer and you hire someone, you pay payroll taxes on their earnings. But if you buy a robot, you can usually write it off right away as a business expense. The tax system nudges you toward replacing people with machines." He says such a tax "would slow the rush away from human labor a little" and fund retraining, and would need targeting so as not to slow beneficial uses in medicine and education. He acknowledges and answers the efficiency objection: critics "point out that it's not optimally efficient in an economic sense, but they're not considering the broader value of work for individuals and society," and "we'll be able to afford a little inefficiency as the price for keeping people employed." He notes he proposed a robot tax years ago to a largely dismissive reception and remains a proponent, while calling it "not the whole solution."
On who should decide
Gates argues explicitly against leaving the response to AI developers: "It is great that some AI companies are proposing solutions to challenges raised by their own technology, but we should not expect them to lead the charge. Some of the issues are outside their area of expertise, and in a democratic society it's not their role to decide these things." He wants solutions developed "through a public democratic process" including elected officials, policymakers, educators, health workers, local officials and community leaders, notes that "local communities are already raising concerns about the energy and water needed for data centers," and warns that without solutions "some groups will push for stopping AI development and deployment altogether." He assigns religious leaders a role and cites Pope Leo XIV's encyclical Magnifica Humanitas by name, saying it "lays a strong foundation for the work that needs to be done" (Magnifica Humanitas).
In the closing section he commits to raising AI and equity with lawmakers in Washington and with leaders abroad, to advocating for the framework, to Gates Foundation work on beneficial use including in Africa, to Breakthrough Energy's use of AI for clean energy, and to writing about AI regularly.
Attribution boundary with the coverage
Three widely quoted lines are from an accompanying Axios interview, not from the memo, and do not appear in its text: that "on the current course and speed, there's a very high chance of a net negative outcome"; that the risks are "all like flashing either yellow or red"; and that "neither party is very knowledgeable or being thoughtful about this" (Source: axios.com). Separately, Axios's account of Human Reserved names child care and jury service as candidate occupations; the strings "child care," "childcare," and "jury" do not appear anywhere in the memo, whose examples are elder care and terminal diagnosis (Source: axios.com). Whether those two examples come from the interview or are a coverage error is not established.
Standing relative to the evidence
The memo is a position, and its central empirical claim — that there will be "far fewer" jobs than exist today absent intervention — is considerably stronger than the measured displacement findings collected at AI Labor Disruption, which remain equivocal. The one study Gates cites for present-tense effect, the Stanford "Canaries in the Coal Mine" paper, he characterises accurately as showing white-collar employment "already being hit modestly," concentrated among young workers in exposed occupations. The projection beyond that is argument rather than measurement.
Provenance
Published on gatesnotes.com on August 26, 2026 under a Bill Gates byline, in eight sections. The page is client-rendered and returned no body to the fallback fetcher used by the morning digest of the same date, so the item initially ran entirely through Axios coverage; the primary was captured with a rendering fetch on August 26, 2026 and verified against first-party domain, byline, publication timestamp, and OpenGraph metadata. Corroborated independently by two Axios articles of the same date.
Relationships
- related: Bill Gates — the author; this is the anchor source for that page.
- related: AI Labor Disruption — the memo's subject; its projection is stronger than the measured evidence collected there.
- related: AI Governance (umbrella), Frontier AI Governance — the two institutional tiers proposed.
- related: AI Public Wealth Fund and Government Equity in AI — a peer mechanism for distributing AI gains, against which the token-and-robot tax sits.
- related: Magnifica Humanitas — Encyclical Letter of Pope Leo XIV (15 May 2026) — cited by name in the memo as a foundation for the work.
- related: Microsoft — the company Gates co-founded and a disclosed financial tie.
- related: AI Companions — the third risk section's subject.