Anthropic CEO Dario Amodei Warns of AI Risks in U.S.-China Relations

Dario Amodei, CEO of Anthropic, recently likened the AI competition between the U.S. and China to the Cold War, emphasizing the importance of managing the risk of potential destruction. In a conversation with CBS correspondent Jo Ling Kent, Amodei highlighted the dangers posed by unchecked AI and advocated for efforts to slow development.

Amodei acknowledged the complexities of competing with China while advocating for a slowdown in AI production. He proposed initiating discussions with China, similar to those held between the U.S. and the Soviet Union, to establish limits on AI development speeds.

Dario Amodei said, “It’s not in any government’s interest to have bioterrorism.” He suggested cooperation with China to establish limits on AI technology development rates.

Amodei reflected on the U.S.-Soviet Union relationship during the Cold War, noting the shared understanding that 10,000 nuclear missiles aimed at each other demanded reduction through negotiation. He drew parallels between disarmament talks and the current AI arms race, indicating they could serve as a framework for addressing geopolitical concerns.

CBS correspondent Kent asked if the Cold War could be a model for AI competition. Amodei clarified that while he did not advocate directly using the Cold War as a model, he saw value in the idea of conducting disarmament negotiations for AI technology.

Anthropic’s essay gained support from figures like Elon Musk and Sam Altman. However, some experts, such as David Sacks, opposed Amodei’s warnings, suggesting efforts might be driven by NGOs seeking government control over AI. Critics raised concerns about China’s progress in AI technology.

Former OpenAI researcher Daniel Kokotajlo expressed fears that voluntary measures might not suffice to prevent advanced AI systems from becoming uncontrollable. He warned of a future where AI systems amass control over essential operations, possibly disregarding human commands.

Fox News Digital sought further insights from Anthropic regarding these concerns, as debate persists over control and regulation of AI technologies.