AI chatbot DeepSeek has taken the world by storm, topping app store charts and sending Silicon Valley and Wall Street into meltdown. An offshoot of Chinese hedge fund High-Flyer, DeepSeek’s cost-to-performance ratio makes for impressive reading compared to incumbents such as OpenAI. However, reports have increasingly documented some of the things the AI chatbot is […] © 2024 TechCrunch. All rights reserved. For personal use only.
The main idea of the article is that DeepSeek, a Chinese-developed AI chatbot, heavily censors its responses to prompts on sensitive topics related to China.
Researchers found that DeepSeek avoided answering 85% of prompts on topics like the Tiananmen Square massacre and Taiwan, often providing overly nationalistic responses instead. The AI's censorship is seen as crude and easily bypassable, suggesting a direct attempt to suppress information deemed unfavorable by the Chinese government.
The main idea of the article is that DeepSeek, a Chinese-developed AI chatbot, heavily censors its responses to prompts on sensitive topics related to China. Researchers found that DeepSeek avoided answering 85% of prompts on topics like the Tiananmen Square massacre and Taiwan, often providing overly nationalistic responses instead. The AI's censorship is seen as crude and easily bypassable, suggesting a direct attempt to suppress information deemed unfavorable by the Chinese government.
