Bill Gates: Agreeing on AI rules will be harder than limiting nuclear weapons
Microsoft co-founder Bill Gates said reaching agreement on international rules for artificial intelligence will be harder than limiting nuclear weapons, and called on the United States to take a leading role in creating them.
Abdulrahman Sadikhli3 min read
Microsoft co-founder Bill Gates has called on the United States to take a leading role in creating international rules for regulating artificial intelligence. According to him, getting states to agree on such rules may prove harder than reaching agreement on limiting nuclear weapons during the Cold War. He said this in an interview with the NBC News program "Meet the Press" on September 27, 2026.
"This is harder," Gates said.
According to him, AI requires international cooperation, since its development affects the security of all humanity. Gates advocated introducing safeguards now and said that control over how AI works does not necessarily have to slow the technological race between the United States and China.
In particular, he proposed building permanent monitoring into systems to track the actions of AI agents and identify dangerous behavior.
Threats and control
In Gates's view, the most immediate threat is not the hypothetical emergence of a superintelligence, but the use of existing technologies by people with malicious intent. With the help of AI, small groups could gain capabilities that were previously available mainly to states, he believes.
Among the potential threats, Gates named large-scale fraud, attacks on energy grids and banking systems, as well as assistance in organizing cyberattacks and biological attacks. According to him, states need to ensure that control systems are in place before major incidents occur.
Gates also commented on proposals to create a so-called "kill switch" — a mechanism that would allow a dangerous AI system to be shut down immediately. He considers such a tool useful but insufficient: in his view, shutting down a system alone is not enough, because it is necessary to understand what exactly it was doing and how it came to dangerous actions.
The debate over regulation in the United States
Against the backdrop of these statements, debates continue in the United States over the need for government regulation of AI. President Donald Trump has previously opposed introducing additional federal restrictions for companies developing AI, linking the development of the technology to competition between the United States and China. Gates, by contrast, believes that security measures in themselves should not deprive the United States of a technological advantage.
Gates also spoke about his foundation's work to expand AI capabilities for languages spoken beyond the English-speaking world. According to him, system errors when working with some African languages can currently be about ten times higher than when working with English. The foundation is partnering with about 60 companies to improve the quality of such systems, including for medical and educational tasks.
Source: news.am
