tiananmen square

4+ articles

AI AI censorship AI governance AI technology Alibaba

U.S. Intensifies Scrutiny of Chinese AI for Political Bias

The U.S. government is conducting evaluations of Chinese AI programs, such as Alibaba's Qwen 3, for ideological biases aligning with Chinese Communist Party narratives. These models, often echoing Beijing's views on topics like the South China Sea, are under investigation for promoting specific political stances. The evaluations may be made public to increase awareness of potential AI-driven ideological influence.

Jul 10

U.S. Intensifies Scrutiny of Chinese AI for Political Bias

Sand AI's Censorship Sparks Debate: Open-Source Meets China's Information Controls

Sand AI, a Chinese AI video startup, launched its open-source video-generating model Magi-1, but sparked controversy by censoring politically sensitive images on its hosted version. The censorship, aligning with Chinese regulations, raises questions about freedom of expression and the authenticity of open-source claims. While the model promises high-quality video generation, its aggressive filtering of topics related to Xi Jinping, Tiananmen Square, Taiwan, and Hong Kong has drawn global attention and lively debate.

Apr 23

Sand AI's Censorship Sparks Debate: Open-Source Meets China's Information Controls

Perplexity AI's Bold Move: Integrating DeepSeek R1 to Bypass Censorship

Perplexity AI, under the leadership of its Desi CEO Aravind Srinivas, has devised a way to leverage the Chinese-developed DeepSeek R1 language model without succumbing to censorship hurdles. By hosting DeepSeek’s open-source architecture on US servers, the company ensures uncensored and private data processing, even tackling infamous queries like Tiananmen Square. The service, offering varied query limits between free and Pro users, stands as a testament to merging Chinese AI tech with Western privacy standards. This marks a significant step in the AI domain, challenging geopolitical norms and instigating debates about data privacy and accessibility.

Jan 30

Perplexity AI's Bold Move: Integrating DeepSeek R1 to Bypass Censorship

ByteDance's Doubao LLM Sparks Controversy on Boox E-Readers with Pro-China Bias

Boox e-readers, powered by ByteDance's Doubao LLM, have ignited a storm of controversy by disseminating pro-China biased content, including denying events like Tiananmen Square. The backlash forced Boox to revert to OpenAI's GPT-3. This incident highlights the potential perils of integrating Chinese AI models in global gadgets.

Jan 17

ByteDance's Doubao LLM Sparks Controversy on Boox E-Readers with Pro-China Bias