In brief
An examination of DeepSeek's A.I. chatbot reveals an intriguing balance between providing information and adhering to China's stringent censorship rules, raising questions about the algorithm's implications globally.
As global interest surges in DeepSeek’s capabilities, there’s a pressing inquiry surrounding its functionality amidst Chinese censorship. To explore this, I engaged the A.I. chatbot, registering with a Chinese phone number and using a local internet connection, thereby subjecting myself to the Great Firewall constraints.The conversation yielded unexpected findings. DeepSeek exhibited flexibility compared to several Chinese applications, frequently providing responses containing terms usually eradicated on domestic social media. Yet, there were scenarios where the A.I. self-censored.** The “thinking” component of the chatbot, which deliberated a response before delivery, allowed for hints of the broader information, observable before the program retracted its statements. Conversely, some discussions echoed the typical experience of online navigation in China, as certain terms remained off-limits. The developers of DeepSeek seem intent on mending any identified censorship loopholes, although they were unavailable for immediate comment. Further tests utilizing software to bypass the Great Firewall revealed little variation in responses, implying that international users might face a similar A.I. experience.Thus far, the tendency of the censored internet has been confined to China’s domestic users. However, if DeepSeek expands its presence internationally, it could inadvertently propagate narratives that align closely with Beijing's priorities across global platforms.
