Sam Altman, Elon Musk back Anthropic CEO Dario Amodei’s warning to slow AI development

Following the remarks of OpenAI's Sam Altman about being open to slowing down the development of AI systems, Anthropic's CEO Dario Amodei has now also called for regulating the pace of AI advancements. In his essay "We Must Pace the Frontier," Amodei argued that rapid progress in AI capabilities is outpacing the industry's capacity to test and control these systems.

While acknowledging AI's potential to drive economic growth, Amodei warned that unchecked AI development could pose severe risks if safety research cannot keep up. His comments come amid heightened competition among AI labs and soaring investments in computing infrastructure. Altman echoed Amodei's perspective on X, agreeing with the need to "pace" the development of frontier models.

He also pledged to involve independent auditors with access comparable to employees in bolstering safety measures for AI development. Musk, who leads SpaceX and xAI, also expressed agreement with Amodei's stance on X. Amodei proposed a three-pronged strategy to curb these risks and promote the "pacing" of AI systems. He suggested embedding external evaluators with employee-like access to frontier AI companies, coordinating safety standards among AI firms in democratic nations, and fostering global collaboration between nations.

Amodei acknowledged the difficulty of achieving coordination between the US and China, emphasizing the need for robust verification mechanisms in any agreement with authoritarian governments. He also advocated for restrictions on exporting advanced AI chips and semiconductor equipment to China, stronger measures against chip smuggling, and enhanced controls on unauthorized model distillation and security against the theft of model weights.