US Agencies Accuse China-Based AI Firms of ‘Malicious’ Copying of American Models

The US government’s accusation against six Chinese AI firms for "malicious" knowledge distillation is a significant development for Asia’s tech sector, particularly for companies like Alibaba and Moonshot AI. This isn't merely about intellectual property theft; it points to a hardening stance on AI data acquisition practices and could lead to new regulatory pressures. The US agencies specifically cited DeepSeek’s organized campaign since late 2024 and Moonshot AI’s broad campaign since mid-2025 as examples of industrial-scale extraction, which they differentiate from legitimate research. For Asian AI companies, the key takeaway is the increased scrutiny on how models are trained and data is sourced. The advisory’s recommendation for US providers to degrade answers and share intelligence suggests a more aggressive defense mechanism against such activities. The US government’s mention of potential sanctions for IP theft, as stated by Bessent in July, indicates that this issue could escalate beyond advisories and impact market access or partnerships for named firms. This situation could force a re-evaluation of data sourcing strategies across the region.