HONG KONG — China's latest artificial intelligence chatbot is trained in President Xi Jinping's doctrine, a stark reminder of the ideological parameters that Chinese AI models must adhere to.
The Chinese cyberspace academy announced earlier this week that the chatbot has been trained in the Xi Jinping idea, a doctrine that promotes 'socialism with Chinese characteristics'.
The chatbot was trained on seven databases, six of which were largely related to information technologies of China's internet watchdog, the Cyberspace Administration of China, or CAC.
Xi Jinping Thought was the seventh database on which the chatbot was trained, according to a report from CAC magazine's WeChat messaging service about the AI model.
The Xi-trained grand language model is the latest attempt by authorities to spread the Chinese leader's ideology and ideas. Students were required to take lessons on Xi Jinping thought in schools, and in 2019 an app called Study Xi Strong Nation was also rolled out so users could learn about his ideologies and take quizzes.
The Xi Jinping Idea – also known as 'Xi Jinping Thought on Socialism with Chinese Characteristics for a New Era' – consists of 14 principles, including ensuring the absolute power of the Chinese Communist Party, strengthening national security and socialist values, as well as improving people's livelihoods and well-being.
The chatbot was primarily created to demonstrate its development and practical applications in cybersecurity and IT research, according to the China Institute of Cybersecurity Affairs, which had launched the AI model for internal use.
According to CAC magazine's WeChat post, the AI model could generate report summaries on topics such as AI development and productivity.
“Users can choose different categories of knowledge bases for intelligent question and answer. The professionalism and authority of the corpus ensure the professionalism of the content generated,” the post said.
The AI model will also be able to create reports, summarize information, and provide translations into Chinese or English for users who access it.
However, it is not clear whether the chatbot trained by Xi will be intended for public use.
The chatbot's unveiling comes as the US and China are locked in a race for AI supremacy.
While San Francisco-based OpenAI's generative AI model ChatGPT is widely seen as one of the most important developments in this field, China also has ambitions to become the world leader in AI by 2030.
China's access to a population of 1.4 billion people also allows the country to collect large amounts of data to train facial recognition and autonomous driving systems.
Technology companies such as Alibaba and Baidu have already rolled out mainly Chinese-language AI models, similar to ChatGPT, for public and commercial use.
However, these AI models tend to be more limited because they must adhere to China's strict censorship rules. This means that Chinese AI models will often not answer politically sensitive questions asked of them.
