American AI strategy expert Joe Weinman said communication channels between China and the United States must be secure and verifiable, given AI's capacity for deception, stressing that cooperation between the two powers is vital, especially as AI-related risks could lead to catastrophic incidents.
China and the U.S. have agreed to establish an AI dialogue to exchange views on AI-related risks and benefits, as well as a communication channel for AI-related incidents. The consensus was one of the key outcomes of the latest summit between Chinese President Xi Jinping and U.S. President Donald Trump during Xi's state visit to the U.S. last month.
Weinman, internationally renowned as an expert on cloud computing and its potential impacts on the global economy, has welcomed the announcement of the dialogue. In an interview with China Global Television Network (CGTN), he discussed how such a channel might work and how the two countries can cooperate while managing competition.
"Channels of communication are critically important to ensure that at all levels, whether it's the leadership level, the research level, military levels, and so forth, that the communications remain open, and that, given that AI has already demonstrated the ability to be deceptive, it's critically important that the communication channels are verifiable and secure, so that they don't present another potential attack vector or risk vector," he said.
While nations and companies can benefit from both competition and cooperation, Weinman stressed that AI cooperation is different because the technology poses grave risks, making cooperation between China and the U.S. particularly important.
"Nations or companies can benefit from developing competitive strategic advantage. They can also benefit by cooperating. And in this case, the nature of cooperation is very different. It's one thing to say that, let's say, we're in the rice-growing business. I own a company that does it and my competitor does and we both want to work out new ways of improving our crop yield so we can cooperate. Here, we're talking about potential existential risk. And so it's particularly important that the two leading nations cooperate when it comes to risks that could create catastrophic incidents, which could be risks to critical infrastructure or could be risks to human life or could be risks much broader than that," the expert said.
China-US communication channels on AI must be secure, verifiable: expert
