China's operational guidance to AI companies requires tracking sensitive keywords violating 'core socialist values,' impacting AI progress by companies like ByteDance, Alibaba, Moonshot, and 01.AI.
Chinese regulator CAC demands AI models to address 20,000-70,000 questions for safety, with 5,000-10,000 questions that shouldn't be answered, half related to political criticism that the model must decline.
China's chatbots, like other AI systems, are restricted from discussing sensitive topics like Tiananmen Square or criticizing President Xi Jinping, resulting in conversation termination for politically sensitive queries.
#generative-ai #chinese-government-regulation #global-ai-race #political-speech-control #technology-restrictions
[
Collection
]
[
|
...
]