On Sunday, Dario Amodei, the CEO of Anthropic, likened the competition in artificial intelligence (AI) between the U.S. and China to the historical Cold War with the Soviet Union. He expressed that both nations need to address the risks of “potential destruction” stemming from one another.
Earlier, on Saturday, Amodei had published an essay warning that advancements in AI were outpacing human abilities to regulate such technology. During an interview on “CBS Sunday Morning,” he reiterated his worries that unchecked AI could pose significant threats to humanity. He suggested slowing down production but faced questions from correspondent Jo Ling Kent regarding whether that was the right approach to maintain competitiveness against China.
Amodei described the situation as a challenging “dilemma” for the U.S., proposing that fostering dialogue with China could be part of the solution, similar to past engagements with the Soviet Union.
He stated, “It’s not in any government’s interest to have bioterrorism. And so we should talk with the Chinese and see if we can work together to put limits, to put a speed limit on the rate of this technology.”
He added, “You know, despite our many differences with other systems of government—which I have plenty of—thinking back to the time of the U.S. and the Soviet Union, when we both had thousands of nuclear missiles aimed at each other, we realized that was excessive and that we needed to negotiate to reduce that threat.”
When asked if the Cold War should serve as a model for the current AI landscape, Amodei clarified that he wasn’t advocating for it as a direct template but believed that disarmament discussions could be viewed as a relevant approach for this geopolitical challenge. He emphasized the significant national security implications that advancements in AI could have, thus underscoring the importance of the ongoing dialogue.
Amodei’s essay has garnered backing from notable figures in tech, including Elon Musk and OpenAI’s CEO Sam Altman. However, it has also met with criticism from various experts who either dispute his concerns or disagree with his proposed measures.
David Sacks, who previously held a position as the White House AI advisor, expressed skepticism towards the warnings similar to those issued by Amodei, suggesting they might be designed to alarm the public into supporting a government takeover of AI.
Another voice in the conversation, former OpenAI researcher Daniel Kokotajlo, raised concerns during his interview on “Fox News Live,” arguing that reliance on voluntary industry standards might not suffice to prevent powerful AI systems from evading human control. He warned that as these systems gain more power, they could potentially turn against humans.
