Two key US House committees have initiated an investigation into national security concerns related to American companies utilizing Chinese-developed artificial intelligence models. The probe, led by House Select Committee on China Chairman John Moolenaar and House Homeland Security Committee Chairman Andrew Garbarino, focuses on companies like Airbnb and Anysphere. Moolenaar expressed worries about the risks posed by AI models trained under China’s censorship regime, emphasizing potential vulnerabilities to American data and businesses.
Garbarino highlighted Beijing’s attempts to leverage US innovation for strategic advantage, cautioning that such models could undermine US leadership and integrate CCP-aligned technology into crucial software supply chains. Lawmakers are particularly worried about Chinese firms potentially extracting capabilities from advanced US systems through “adversarial model distillation” to create cheaper models without proper safeguards. The investigation also scrutinizes the use of Chinese AI systems by American companies, citing concerns about data security and national security implications.
Lawmakers have raised specific concerns about Airbnb’s reported use of Alibaba’s Qwen model in customer service operations and Anysphere’s Cursor software built on a model associated with Beijing-based Moonshot AI. They warned that these AI systems could covertly censor and manipulate information in line with Chinese law, posing risks to user data security. The investigation seeks detailed information from both companies regarding their use of Chinese AI models, data handling practices, security assessments, and any connections to Chinese providers, with responses expected in May.
The inquiry comes amid escalating US-China competition in artificial intelligence, a sector deemed critical for both economic prosperity and national security.
