In brief
As China’s internet censorship remains a significant barrier, DeepSeek’s AI chatbot presents both a challenge and an intrigue. This article explores an insider’s experience engaging with the bot and how it operates under strict regulations, offering insight into its dual role of suppression and information sharing.
As the global community continues to analyze DeepSeek and its implications in the AI landscape, a key question emerges: How does this Chinese-developed chatbot withstand the challenges of extensive censorship? To find answers, I accessed DeepSeek’s A.I. chatbot using a Chinese phone number and internet connection, both subjects to the restrictive Great Firewall that blocks numerous international sites, including Google and The New York Times.My interactions yielded unexpected results. In some moments, DeepSeek exhibited lesser censorship than many domestic platforms, supplying responses laden with keywords that would typically be swiftly removed from Chinese social media. However, the chatbot demonstrated self-censorship at times, ceasing responses midway or deleting parts of its answers. Unique to its design, DeepSeek’s “thinking” capability allowed it to ponder its answers before responding, providing insights that could parallel uncensored information available outside China—if one was attentive enough before the content was removed.Conversely, some phrases remained prohibited, reflecting the overall patterns of online navigation within the country. The developers appear to be working rapidly to identify and mend loopholes in its censorship protocols. When I conducted tests using a VPN to bypass the firewall, the responses remained mostly unchanged, indicating that users across different geographic locations may experience the same functionality, thus revealing a concerning possibility: should DeepSeek expand its influence internationally, it might unintentionally disseminate narratives aligned with Beijing’s policies.In summary, DeepSeek's chatbot illustrates an ongoing dance between information access and censorship, emblematic of broader trends in Chinese digital policy and its potential global implications.
