Bill Gates warns of AI’s unequal impact and calls for urgent safety measures

In an interview with The Ezra Klein Show podcast from The New York Times, published on Tuesday, 29, Microsoft co-founder Bill Gates argued that the effects of artificial intelligence on the labor market will not be uniform across the world. He pointed out that wealthy countries will feel the changes first, followed by middle-income economies—such as Brazil, China, Vietnam, and Indonesia—and then by poorer nations.

Gates compared the distinct realities of middle-income and low-income countries, citing differences in child mortality and the availability of doctors. “There are middle-income countries, such as China, Brazil, Vietnam, and Indonesia, where economies are growing and the child mortality rate is at most three times lower than in the United States. On the other hand, there are low-income countries where 15 times more children die before the age of 5, and there are no doctors,” said Gates. “You live your whole life without ever meeting a doctor. You try to figure out what seeds to plant, without ever having anyone to advise you on what to do,” he said.

The businessman believes that treating the Global South as a homogeneous bloc prevents understanding these inequalities. “Treating the Global South as one uniform thing does not allow you to understand what it’s like to live there,” said Gates. “We therefore need a little more nuance. In low-income countries, AI will be overwhelmingly positive and should be promoted as quickly as possible,” he said.

Despite his optimism about AI’s potential to solve basic problems in health, education, and agriculture, Gates warned that the replacement of cognitive tasks could leave many workers without new occupations. “AI is a tool for good, helping people who need it most, with very basic human problems,” he said. “Human cognition has become the scarce element. It’s always a question of scarce resources. When you replace human cognition as the scarce resource—in fact, you surpass it in terms of working 24 hours a day, reading more, knowing all the languages—there’s no scarcity in what you’re progressing,” he said.

Risks may precede benefits

Gates also noted that, without adequate controls, the dangers of AI could arise before its benefits materialize. He rejected the idea that humanity is doomed, but stressed that the automation capability could be used for cyber or biological attacks before regulations and tests are in place. “The control problem is serious, but the imminent risk is not recursive self-improvement. The imminent risk is that these are the most powerful tools ever,” he said.

In contrast to high-risk technologies such as nuclear weapons, Gates noted that AI systems are not developed primarily with government funding. “It’s not like rockets or nuclear weapons,” he says. This distinction, he said, contributes to governments’ lack of understanding of the dangers involved.

To mitigate these threats, he advocates the immediate implementation of safeguards and a mandatory oversight system. “This oversight layer is urgently needed,” he says. Gates argues that waiting for major incidents to act would allow AI capabilities to advance in dangerous directions. “That’s a huge mistake,” he says.

An era of conditional abundance

The Microsoft co-founder still envisions a future of abundance, in which AI can reduce costs in health, food, and housing, but only if humanity maintains control over the systems and manages to limit damage during the transition. “The good things, if we’re not careful, come more slowly than the biological and cyber risks,” Gates said.

He proposes that countries adopt contributions on automated labor, reserve essential activities for humans, and establish mandatory standards for advanced models. These measures, he said, are essential to ensure that the benefits of AI are distributed fairly and that risks are contained.

In summary, Bill Gates argues that public policies and international dialogue must recognize disparities between economies and act proactively to prevent the dangers of artificial intelligence from outweighing its promises.