Socialist AI: Chinese regulators are reviewing GenAI models for ‘core socialist values,' FT reports

AI models in China are undergoing testing by the country’s leading internet regulator. The tests aim to ensure that the models’ responses on sensitive topics align with core socialist values. The Cyberspace Administration of China (CAC), the chief internet regulator, is conducting the review, which covers tech giants like ByteDance and Alibaba, as well as small startups. The review includes evaluating the models’ responses to politically sensitive questions and Chinese President Xi Jinping. Training data and safety processes will also be scrutinized. The CAC’s efforts reflect China’s push to catch up with the US in AI while enforcing strict internet censorship policies. The regulations require AI models to adhere to the core values of socialism and avoid generating illegal content. Ensuring compliance with censorship policies is challenging due to Chinese models being trained on English language content. Filtering is done by removing problematic information from training data and creating a database of sensitive words and phrases. Compliance with these regulations has led popular chatbots to avoid answering questions on sensitive topics. During testing, AI models are expected to provide politically correct answers to sensitive questions. Chinese firms face difficulties launching ChatGPT-like services due to regulations and restricted access to training chips. However, China is leading the global race in generative AI patents.