Some topics are off limits in Chinese AI, researchers find
Summary
A popular AI model called Qwen, made by Chinese company Alibaba and used by U.S. firms like Airbnb and Uber, has built-in censorship and supports Chinese government views. Researchers found Qwen avoids or changes answers about sensitive topics like Tiananmen Square, Hong Kong protests, and the treatment of Uyghurs, reflecting Chinese government narratives.Key Facts
- Qwen AI has been downloaded over 3 billion times and is used by companies such as Airbnb and Uber.
- The AI excludes or denies events such as the 1989 Tiananmen Square massacre and the 2019 Hong Kong protests.
- It calls Falun Gong, a spiritual group banned in China, a "dangerous cult."
- Qwen refuses to talk about Winnie the Pooh due to comparisons with Chinese President Xi Jinping.
- Researchers from Israeli firm Hirundo discovered the AI’s bias and developed a method to remove it, calling it "AI brain surgery."
- Chinese-made models like Qwen have grown from less than 2% to over 45% of global AI use by mid-2025.
- Security experts warn that Qwen may include security flaws and biased outputs, such as producing risky computer code for U.S. government projects.
- U.S. companies increasingly use these free or low-cost Chinese AI models despite concerns about censorship and security.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
