A U.S. House panel states that China's DeepSeek AI is a 'profound threat' to national security

DeepSeek AI's China ties threaten US, urging export curbs.

: The House Select Committee scrutinizes DeepSeek AI's connections with China, highlighting the risks posed by its data practices. DeepSeek allegedly uses 'illegal' distillation from OpenAI’s models, escalating concerns over national security. The report coincides with a Trump administration move to restrict Nvidia chip exports, seen as a strategic challenge. Microsoft and OpenAI accuse of exploiting U.S. technologies to gain competitive advantages through distillation methods.

The U.S. House Select Committee on China issued a report labeling DeepSeek AI as a 'profound threat' to national security, emphasizing its close ties with the Chinese government, which allegedly uses DeepSeek to collect and send user data back to China. Liang Wenfeng, DeepSeek's founder, is reported to have connections to a hedge fund tied to state-linked labs.

The report suggests prohibiting federal use of Chinese AI models and imposing export restrictions on AI technologies to China. Nvidia, a chipmaker, experienced a setback as the Trump administration expanded existing restrictions on chip exports to China, previously aimed at low-end chips, potentially stimulating China's self-reliance initiatives.

OpenAI and Microsoft have claimed DeepSeek employed a distillation technique, learning from OpenAI’s ChatGPT model outputs without consent, to enhance their AI models, violating intellectual property. The distillation process involves an AI model learning from another's reasoning steps to swiftly advance its capabilities.

Critics argue the U.S. export restrictions could inadvertently encourage Chinese innovation, as companies like Huawei have previously accelerated internal chip production following Western cutoffs. Historical precedents suggest high-tech restrictions might not hinder China's technological growth permanently.

DeepSeek's potential for geopolitical influence concerns U.S. lawmakers, given the control AI developers can exert over model outputs and potential bias. While China progresses its AI landscape, America strives to safeguard its technological advantages through stringent export policies.

Sources: Thomas Maxwell, Bloomberg

Security