Chinese AI Chatbot DeepSeek Censors Itself in Real-Time
Users of the Chinese AI chatbot DeepSeek report witnessing its self-censorship in real-time, as it begins thoughtful responses to sensitive questions, only to delete them abruptly and replace them with vague messages like, “Let’s chat about coding instead.”
While regulated versions of DeepSeek comply with Chinese government restrictions, open-source versions of its R1 model available on platforms like Hugging Face provide uncensored answers, discussing topics like Tiananmen Square and Taiwan’s independence more freely.
This duality highlights the tension between DeepSeek’s technological achievements and the political constraints governing its use, raising questions about how censorship and AI ethics will evolve in the global market.
Source: The Guardian
For more details, read the full article here.