Bill Gates' Latest Interview: AI Is More Dangerous Than Nuclear Weapons, and Could Be More Useful Than Nuclear Energy
In an interview with the Vatican media outlet Vatican News, Bill Gates judged that AI could be more dangerous than nuclear weapons and could also be more beneficial than nuclear energy, and said that the tech industry has reached a consensus that self-regulation alone is not enough, while government policymakers lack the expertise. He mentioned that the foundation has committed to investing $1 billion over the next two years to promote AI in healthcare, education, and agriculture so that it benefits underserved populations. He also believes that society needs to discuss in advance how to tax human labor and automation, and whether child care, judges, jurors, and some medical procedures should be reserved for humans.
Body · AI translation
[[**Core Summary Body**
**Key Points**
**1. AI carries both enormous risks and benefits**
**2. Industry consensus is ahead; government capacity lags behind**
**3. The foundation is investing $1 billion to promote inclusive AI**
**4. Near-term misuse risks are more urgent than AI "awakening"**
**5. Society needs to decide which jobs should be reserved for humans**
Even if AI can perform certain tasks, child care, judges, jurors, and parts of the medical process may still need to be left to humans, because this involves responsibility, trust, and public values, and taxation and employment arrangements need to be discussed in advance. 2026-10-15 15:27 Bill Gates' latest interview: AI is more dangerous than nuclear weapons, and may also be more useful than nuclear energy. Quick overview by Leng Mou
Bill Gates has recently been talking about artificial intelligence more and more frequently, and his warnings have become increasingly specific. On October 11, on CNN's "State of the Union," he turned his attention to the more imminent risk of cyberattacks. Two days later, the Vatican media outlet Vatican News published an interview with him by reporter Roberto Cetera. This time, he offered a striking judgment: AI is more dangerous than nuclear weapons, but the benefits it may bring also exceed those of nuclear energy.
On October 13, Vatican News published an interview with Gates by reporter Roberto Cetera. The most striking statement was: AI may be more dangerous than nuclear weapons, and may also be more beneficial than nuclear energy. These two halves must be understood together. Gates is not talking about a machine destined to destroy the world, but about a technology whose potential benefits and potential harms are both extraordinarily large.
The tech industry has already recognized the risks, but governments have not yet caught up
In Gates' view, the notable change is that technology companies and industry insiders who usually compete with one another and hold different views are forming a common understanding: AI cannot rely on industry self-regulation alone. The problem is that many government policymakers still lack enough expertise to understand model capabilities and related risks.
But he is equally dissatisfied with another imbalance: people are not making full use of the social benefits AI can already provide. Gates mentioned the foundation's Goalkeepers report: AI may narrow the wealth gap, or it may widen it. Rural doctors can use AI and smartphones to improve ultrasound examinations; students can receive tutoring better suited to their own learning styles; farmers can receive advice combining local weather and soil conditions. These are not an abstract utopia far in the future, but real-world applications he hopes to prioritize.
The problem is that the market will not automatically deliver the best tools to the people who need them most. Those with greater ability to pay may get services first, while regions most lacking in medical and educational resources continue to fall behind. Gates said the foundation has committed to investing $1 billion over the next two years to promote AI benefits in health care, education, agriculture, and other fields for populations underserved by the market.
Gates lived through the personal computer and internet revolutions, but he believes past experience cannot simply be applied this time. What makes AI special is that humanity has created an intelligence that can match or even surpass humans in many cognitive tasks. He describes it as a change closer to an "evolutionary event" rather than the next generation of personal computers.
He distinguishes two phases. The current one is a transition period: the broad benefits of AI have not yet truly materialized, but the impact on jobs and social structures has already begun. In wealthy countries, where knowledge work accounts for a higher share, the impact may be more pronounced; in low-income regions lacking doctors, teachers, and other services, AI may first fill long-standing supply gaps.
Even so, language barriers remain for inclusive technology. Many models are trained mainly on English-language materials and cannot serve billions of people who use other languages with equal quality. The expansion of technological capability does not mean everyone can obtain that capability equally.
If safety and accessibility can be properly handled in the future, Gates envisions that AI and humanoid robots may have the opportunity to take on a large amount of labor related to basic needs such as food and housing, and society would thereby enter a more affluent stage. At that point, how people arrange their time and find meaning will become important questions. But he stresses that what truly needs to be addressed first is the transition over the next few years: how to avoid serious biosecurity and cyberattack incidents, and how to prevent large numbers of workers and some countries from being left behind.
Speaking of the responsibility of political leaders, Gates made a striking comparison between nuclear weapons and nuclear energy. He believes that implementing basic safety checks need not significantly slow technological progress, but could reduce a large amount of risk. Governments should bring in people who understand AI, cooperate with companies, and cooperate with other countries. No country wants to suffer bioterrorism or have critical infrastructure paralyzed by cyberattacks.
What worries him is precisely that policy responses are too slow: society may have to wait until a major cyber or biosecurity incident occurs before it begins to take the issue seriously. Rather than limiting the debate to whether AI will become conscious, Gates is more focused on the misuse risks and governance responsibilities that can already be discussed.
In another interview on CNN's "State of the Union" on October 11, Gates further emphasized the near-term risk of using AI to carry out cyberattacks. This is a supplementary point from a separate interview, not a quote from the October 13 interview. An earlier interview on NBC's "Meet the Press" on September 27 also touched on safety guarantees and regulation; a transcript page that appeared online on October 12 does not mean the interview took place that day.
Even if machines do it better, which jobs should still be done by humans?
A longer-term question Gates raises is this: society needs to decide how to tax human labor and automation, and which occupations should be deliberately reserved for humans. He cited child care, judges and jurors, and certain parts of medical work.
The value of these occupations is not only the accuracy of completing tasks. Relationships between people, public trust, and responsibility itself are also part of the work. Even if AI has the ability to complete a task, it does not follow that society must remove humans from it. The point here is not that Gates has already proposed a specific "robot tax" bill or a list of occupations where AI should be banned—he has not—but that he is asking society to discuss this choice in advance.
At the end of the interview, when discussing Pope Leo XIV's remarks on AI and human value, Gates said what he agrees with is not stopping AI, but actively safeguarding human value, and he welcomed religious groups to join the discussion. Asked whether his departure from Microsoft's day-to-day management was out of concern about AI, he also clearly denied it: that was a transition to philanthropy arranged gradually more than a decade ago and cannot be explained as being caused by recent AI risks.
Further thought: when efficiency is no longer scarce, who decides where people belong?
A young person entering the workforce often gains experience by organizing materials, writing first drafts, and completing basic analysis. If more of these tasks are handed to AI, companies may reduce entry-level positions, or they may redesign how people are trained. Importantly, this is not Gates's quantitative forecast of employment levels, but an industry question worth continuing to track: if newcomers lack opportunities to practice, where will the next generation of professionals grow from?
Another kind of inequality may also emerge: wealthy families can continue to buy the time of human teachers, doctors, and caregivers, while other families mainly use lower-cost automated services. At that point, will the "inclusiveness" brought by AI unintentionally create a stratification of services? This is precisely the question Gates leaves society with when he emphasizes technology accessibility, human value, and tax arrangements.
Technology can answer whether a job can be automated, but it cannot by itself decide whether a job should be automated. The latter question involves responsibility, dignity, and how wealth is distributed. Compared with predicting how smart machines will ultimately become, this may be what is more worth discussing seriously in Gates's latest exclusive interview.
Source:
Has published 33 articles on Huxiu. This content may not be reproduced without permission. For authorization, please contact hezuo@huxiu.com. Everyone who is changing and wants to change the world is on the Huxiu APP[ ]
AI Trends