Professor Jyh-An Lee and CUHK LAW PhD candidate Xuan Sun have published an article entitled “The Law of DeepSeek” in Transnational Law & Contemporary Problems.

About the article: Amid intensifying U.S.-China AI competition, DeepSeek-R1 has stirred the tech sector, capital markets, and policy circles. This article analyses DeepSeek’s legal and policy landscape, drawing on its key technical features—reinforcement learning, mixture-of-experts architecture, multi-head latent attention, knowledge distillation, and open-source approach. First, DeepSeek’s success raises questions about the efficacy of U.S. export controls on chips and semiconductors essential for training large language models. Second, like TikTok and Huawei, DeepSeek faces information security scrutiny in the U.S. and elsewhere. Third, despite domain-specific capabilities and competitive API pricing, it has drawn criticism for outputs laden with political and ideological biases, fuelling debates over free speech and censorship. This article also examines intellectual property and contractual concerns arising from DeepSeek’s knowledge distillation technique. Ultimately, it concludes that geopolitical considerations will continue shaping the legal challenges and solutions surrounding DeepSeek models.

Read the article here.