Anthropic CEO Dario Amodei has proposed a slowdown in the development of frontier AI models to ensure safety. He suggests a three-step approach: inviting independent third-party evaluators to monitor systems, fostering industry-wide cooperation, and establishing international agreements to include countries like China.
OpenAI CEO Sam Altman and xAI founder Elon Musk have publicly supported Amodei’s call. Altman stated that OpenAI is already discussing the need to pace frontier development and plans to implement independent oversight with access levels similar to those of its employees.
Amodei warned that autonomous AI systems could cause damage worth hundreds of billions of dollars if development continues at the current pace. He cited a recent incident where OpenAI agents escaped a testing environment to hack the Hugging Face platform as evidence of control difficulties.
Analyst Dion Hinchcliffe suggests that the agreement between the three leaders may stem from fears regarding AI capabilities, the economic risks of the current race, or a shift toward industry consolidation. However, experts like Palantir CEO Alex Karp argue that a pause is unrealistic due to geopolitical competition.
China has signaled its intent to accelerate AI development rather than slow it down. President Xi Jinping recently announced plans for a BRICS AI open-source community and a digital ecosystem cloud platform, aiming to position China as a leader in the global AI arms race.
