Anthropic cofounder says Chinese models trail West by 6-12 months

According to ETtech, Jack Clark, Anthropic's co-founder and head of policy and public affairs, shared his assessment of China's AI progress in a recent interview. He said Chinese models currently trail Western ones by roughly six months to a year, adding that the models are too unintelligent, which is why runaway agent swarms have not yet emerged. He also warned that in a few months China will confront the same risks the West has already experienced. Only this single set of remarks has surfaced in publicly available reporting so far. Clark did not name specific models or benchmarks to back up the gap, nor did he provide evidence for the claim that Chinese developers will soon face the same safety problems Western teams have already encountered.