The large language model was trained on the political philosophy of the Chinese leader known as "Xi Jinping's Ideas on Socialism," as well as other official literature provided by the Chinese Cyberspace Administration.
"The expertise and authority of the resource guarantee the professionalism of the content," the message about the new LLM, reviewed by the publication Financial Times, states.
Currently, the model is being used at the local internet regulator's research center, but it is planned to be released for broader use in the future. It is reported that the chatbot can answer questions, create reports, summarize information, and translate between Chinese and English.
Over a dozen books have been published in the name of Xi, and they usually take a central place at book fairs in the country. Popular news applications from companies like Tencent or NetEase also reserve spaces at the top of channels for articles from official media, mostly with images of the Chinese leader.
Officials have also demanded that students aged 10 study Xi Jinping's political philosophy and have launched the Study Xi Strong Nation app for educating and testing the knowledge of about 100 million members of the Communist Party in China.
Other developers of generative chatbots are also required to comply with licensing conditions that prescribe "embodying the core socialist values" in the content. Additionally, the Chinese Cybersecurity Association (a non-profit organization affiliated with the authorities) has created a public database with 100 million records of "high-quality and reliable data" for training artificial intelligence models (mostly based on government decrees and political documents, reports from state media, and other official publications).
One of the dozens of text documents in the data package contains 86,314 mentions of Xi Jinping.
"Let's unite even closer around the Party Central Committee with Comrade Xi Jinping at its core," one line states.
Comments (0)
There are no comments for now