At the Financial Times-hosted AI Future Summit, Microsoft co-founder Bill Gates did not mince words. He called the capabilities demonstrated by AI models "astonishing," and said their potential for both good and harm "must not be underestimated." What alarmed him most was malicious actors using AI to create biological weapons or tear open cybersecurity gaps across sectors—while some "open" models are already freely available, and their safety guardrails and monitoring can be stripped away directly. He said bluntly that if AI is involved in a biosecurity incident, the scale of damage "could be far worse than COVID." This is not the first time Gates has sounded the alarm; he had previously warned that AI could claim a billion lives and disrupt employment on a massive scale.
Taking Aim at the White House Gates blamed Washington's absence for the international community's weak response. He said efforts to regulate and monitor AI are being hindered by White House inaction: "Unless the United States participates, action means little; we are stuck in a stalemate." This was clearly aimed at Trump—who refuses to strengthen government regulation, describes drafting G20 AI guidelines as a "globalist conspiracy," and instead bets on industry self-regulation. Gates hit back on the spot: self-regulation simply does not work, and "when multiple companies are involved, the only way to solve the problem falls on governments."
The Three Remedies He Prescribed Gates suggested that the United States set unified evaluation benchmarks for models' internal safeguards and monitoring, lead global cooperation to patch cybersecurity vulnerabilities, and build a monitoring network covering "air, wastewater, and blood" to catch AI-fueled infectious diseases as early as possible. He predicted that OpenAI and Anthropic will both launch new models this year, with "two major releases" in 2027, and concerns about "models themselves going out of control" may be tied to these versions, urgently requiring third-party audits. At the same time, he called for accelerating the release of AI's benefits, such as advancing drug development and eliminating excessively long hospital waiting lists, and explicitly stated support for taxing the super-rich. He also praised Anthropic for acknowledging in last week's IPO prospectus AI's "catastrophic or existential risks," calling such transparency "yet another example."
