US accuses Chinese firms of systematically extracting American AI technology

US authorities have levelled fresh accusations against major Chinese technology companies, claiming they have engaged in what officials describe as systematic efforts to tap into American artificial intelligence models and repurpose their capabilities. The allegations focus on Alibaba and DeepSeek, two prominent players in China's AI sector, and point to techniques employed throughout 2024 and beyond.

According to statements from American agencies, the companies have utilised a method called distillation to access US-developed AI models and extract their underlying information. This technique, which involves querying advanced models to gather training data, has allowed the Chinese firms to operate at an industrial scale, effectively bootstrapping their own AI systems using insights derived from American technology.

Model distillation works by using established AI systems as teachers to train new models, creating a pathway for technological knowledge transfer. The practice raises significant concerns about intellectual property protection in the rapidly evolving artificial intelligence sector, where the underlying models represent enormous investments in research and development.

The allegations reflect mounting tensions between Washington and Beijing over AI technology access and development. The US has increasingly scrutinised Chinese access to advanced computing capabilities and sophisticated AI systems, viewing them as potential national security concerns. Indian technology firms and policymakers monitoring the sector will be watching closely as these international disputes over AI governance continue to develop.