Bill Gates warns of AI risks and calls for US-led global rules

The most urgent threat is its malicious use

In an interview aired on NBC’s Meet the Press on Sunday, 27, Microsoft co-founder Bill Gates said the biggest current danger of artificial intelligence is its use by malicious actors for fraud, cyberattacks, disruption of power grids, and bioterrorism.

Security proposals: monitoring and emergency shutdown button

Gates argues that AI labs should be required to install built-in monitoring systems and an emergency shutdown mechanism, known as a “kill switch.” He says the button alone is not enough, and it is also necessary to record system activities to enable subsequent investigations.

Global regulation likened to the nuclear arms race

The businessman suggests that the United States should lead the creation of a global regulatory framework for AI, warning that reaching an agreement would be more difficult than Cold War negotiations on nuclear weapons. He disagrees with President Donald Trump’s position, who rejected any federal regulation of AI companies and said the US should maintain its advantage over China.

Dialogue between Washington and Beijing

Following a meeting between Trump and Chinese President Xi Jinping on Thursday, 24, the White House announced a dialogue channel between the two countries to discuss the risks and benefits of AI, with a new round scheduled for November. Congressman Ro Khanna proposed the formation of a joint working group with technical experts from both countries.

Industry pressure to slow down development

Recently, Jacob Coxon, a former researcher at Anthropic, warned that AI builders believe the technology could kill everyone by the end of the decade. Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman agreed that it is necessary to slow down development to prevent a superintelligence that surpasses human reasoning, but only enough to maintain leadership over China.

Gates Foundation initiative for African languages

The Gates Foundation is working with 60 AI companies to improve the accuracy of models in African languages, with the goal that within five years, all languages spoken by at least one million people will achieve a quality comparable to that of English.