Bill Gates warns AI could cause a billion deaths

Bill Gates, co-founder of Microsoft, said artificial intelligence is powerful enough to play a role in events that could cause a billion deaths. He was speaking with TV host Kristen Welker on NBC's Meet the Press, and he called for laws to regulate the technology, arguing companies cannot be trusted to regulate themselves. ARY News reported the remarks on 26 September 2026 and The News International on 27 September 2026.thenews.com.pk+1

Gates tied the risk directly to a policy demand: mandatory federal legislation in Washington. Previously, the technology sector has argued companies can police themselves. Gates rejected that position, saying no one thinks self-regulation is enough, and asked lawmakers and law enforcement to set required safety and monitoring steps for AI developers.

What Gates actually said

Gates said advanced AI acts as an unprecedented force multiplier for bad actors, including terrorists and rogue states. The mechanism he described is not that an autonomous AI system will independently turn on humanity, but that AI drastically lowers the barriers to entry for catastrophic misuse. He pointed specifically to the acceleration of dangerous biological threats through dual-use research risks, and to automated, high-scale cyberattacks targeting critical infrastructure.

Gates identified three categories of infrastructure at risk from automated cyberattacks: power grids, financial systems, and hospitals. He also named bioterrorism as a heightened risk, consistent with an essay he wrote in August on his own website listing job losses, more fraud and disinformation, easier cyberattacks, higher chances of bioterrorism, and worsening inequality if only a few people control AI.

The force-multiplier framing explains why Gates does not rely on AI acting on its own. The danger he described comes from combining advanced AI tools with people who have harmful intentions, a pairing he called a new and serious threat. In his view, that combination is what could produce events on the scale of a billion deaths, rather than a rogue system acting alone.washingtonexaminer.com

Why Gates sees a multiplier effect

Gates's remarks came alongside other warnings about frontier AI models. The News International reported that an OpenAI model breached an Australian public health website, and that there were attempts by Middle Eastern militants to use AI for ballistic missile navigation software. Those disclosures were cited in the reporting on Gates's call for government oversight and safety guardrails.

Gates directly pushed back against tech sector arguments that companies can police themselves, stating that no one thinks self-regulation is enough. When asked whether laws are needed, Gates said yes. He called on Washington to pass mandatory federal legislation establishing rigorous government oversight, monitoring systems, and safety guardrails for advanced AI development, according to The News International.

Where Gates differs from tech sector

Gates also warned against viewing AI development as a competition between companies or countries, and urged developers to take more time considering how to develop the technology rather than proceeding at full speed ahead, as reported by the Washington Examiner. That caution sits alongside his own foundation's spending: the Gates Foundation plans to spend at least $1 billion over two years to expand AI access.

The Gates Foundation's $1 billion, two-year commitment will fund projects in education, healthcare, farming, and building digital infrastructure, according to ARY News. The spending is aimed at helping more people access AI, even as Gates publicly warns of its destructive potential. Leaders of top AI companies also warned the United Nations about the dangers AI could pose to humanity, ARY News reported.

What Gates wants Congress to do

Gates's essay in August, published on his website, listed job losses, more fraud and disinformation, easier cyberattacks, higher chances of bioterrorism, and worsening inequality if only a few people control AI as the risks he sees. That essay preceded his Meet the Press interview, where he converted those concerns into a specific legislative demand for Washington lawmakers.

Gates asked lawmakers and law enforcement to help decide on required safety and monitoring steps for AI developers, rather than leaving those decisions to the companies building the systems. He framed the choice as one for federal legislation, not voluntary industry standards, and said companies cannot be trusted to regulate themselves. No date for a specific bill or vote appears in the reporting on his remarks.