In brief
An examination of DeepSeek’s A.I. chatbot shows it struggles with Chinese censorship, often self-censoring in response to sensitive topics. Although it offers answers beyond standard domestic platforms, it still mirrors the restrictions of China's Great Firewall.
As the world races to comprehend the capabilities and implications of DeepSeek, a Chinese-developed A.I. chatbot, a pressing question arises: How does it contend with the stringent censorship imposed by the Chinese government? I sought to uncover this during a conversation with the bot, having registered using a domestic Chinese phone number and internet connection, thus placing myself under the limits of China’s Great Firewall.Surprisingly, DeepSeek exhibited a paradoxical nature when handling sensitive topics. On occasions, it provided answers that would typically be expeditiously deleted on traditional Chinese platforms, indicating a lesser degree of immediate censorship. However, it also self-censored midway through responses. Employing a unique “thinking” feature, the bot occasionally presented information insightfully before retracting it, allowing perceptive users to glean unfiltered insights.In separate tests using firewall circumvention tools, I noted that the responses remained consistent, implying a similar experience for users both within and outside of China. This raises concerns that if DeepSeek continues to expand its influence internationally, it might effectively disseminate Beijing’s narrative globally, shifting the nature of internet discourse beyond its borders.While it indeed reflects the ongoing issues of internet usage in China—where certain words remain unacceptable—the developers of DeepSeek seem engaged in an ongoing struggle to patch the censorship gaps. With increasing accessibility, the chatbot could become a vessel for controlling information, further intertwining technology and state narratives, thus posing a new challenge for global internet freedom.
