China's top AI researchers fear a 'Chernobyl moment' from the US arms race
experts at a beijing conference warned that frontier ai risks outweigh rivalry, calling for us-china cooperation to prevent a 'chernobyl moment' from agentic models.
a beijing conference featuring figures like andrew barto highlighted growing anxiety among chinese researchers that unchecked ai development could lead to a 'chernobyl moment'. mit's stephen casper and shanghai jiao tong university's lin yun argued that despite us chip sanctions and rivalry, cooperation on safety standards is necessary to handle systemic risks like recursive self-improvement and automated cyberattacks. the experts believe that without shared protocols, the proliferation of advanced agentic models could cause global instability, suggesting that defense mechanisms must be developed alongside international coordination.