US Frontier AI Companies Warn Over Distillation Attacks; China Threatens Countermeasures

US AI developers and authorities are concerned about 'distillation attacks'—training models on outputs from other frontier models—allowing China and Russia to create similar systems more cheaply. China rejected the accusations and promised 'countermeasures' if the US uses this as a pretext to constrain its AI. The topic may be raised at the US-China leaders' meeting on September 24.

• Distillation lets smaller models train on frontier model outputs more cheaply
• China rejected accusations, threatened countermeasures against containment
• Anthropic blocked accounts and IPs for attempts to extract Claude reasoning
• Topic may be discussed at US-China leaders' meeting on September 24