The CEO of an AI development platform says that the development of high-performance AI models is concentrated in China, “there is a very high possibility that political censorship will occur” – GIGAZINE

As AI development accelerates globally, Chinese open source AI models are showing superior performance in various AI tasks such as coding and inference. However, the platform, also known as the “GitHub of AI development”,Hugging FaceClement Delang, CEO of , pointed out that developing based on China’s high-performance open source AI may lead to unexpected glitches due to censorship.

Hugging Face CEO has concerns about Chinese open source AI models | TechCrunch
https://techcrunch.com/2024/12/03/huggingface-ceo-has-concerns-about-chinese-open-source-ai-models/

Western AI Firms Are Worried About ‘Censorship’ In China’s AI Models | CCN.com
https://www.ccn.com/news/technology/western-ai-firms-censorship-china-model/

Mr. DeLang is a member of LinkedIn, a business-specific SNS.PostI made six predictions regarding AI in 2025. The predictions included developments in AI, such as “the first large-scale AI protests” and “at least 100,000 personal AI robots to be pre-ordered.” As a result, China will begin to lead the global AI race in 2025,” he said, referring to China’s growth.

In fact, as of 2024, Chinese AI companies will be at the forefront of AI development. in 2023Artificial general intelligence (AGI)DeepSeek was established with the aim of developing the large-scale language model “DeepSeek-R1-Lite-Preview,” which is said to be comparable to, and in some cases even superior to, OpenAI’s inference model “o1-preview.” ” was announced in November 2024. R1-Lite-Preview is also scheduled to release an open source version and API, and may become the basis for future AI development. Other Chinese AI companies such as Alibaba and Moonshot AI are making rapid progress in the AI ​​field by focusing on open source development.

Chinese AI company DeepSeek releases inference AI model “DeepSeek-R1-Lite-Preview” comparable to OpenAI o1 and plans to open source it – GIGAZINE

However, according to a test conducted by technology media TechCrunch, R1-Lite-Preview refused to answer questions about President Xi Jinping, Tiananmen Square, and the geopolitical implications if China invaded Taiwan. . TechCrunch pointed out that AI models developed in China are likely to extensively censor “topics that the Chinese government deems sensitive.”

DeLang expressed similar concerns on his podcast. “If you ask a chatbot based on a Chinese AI model about the Tiananmen Square incident, it won’t respond in the same way as a system developed in France or the United States,” DeLang said. “If a country like China reaches an overwhelmingly high level of AI, it will be able to spread certain aspects of its culture that Western countries would not want spread.”

Other major AI companies have also expressed opposition to China’s AI model. OpenAI researcher Steven Heidel posted a meme image on X criticizing China’s AI censorship in response to DeepSeek’s failure to respond to the Tiananmen incident. In addition, OpenAI CEO Sam Altman expressed concern in 2023, “We are very concerned about authoritarian governments developing AI models.”

According to Mr. DeLang, the problem is that top-class open source models are concentrated in China and other countries because AI models include a considerable amount of the ideology of the country in which they were developed. “It’s important that AI is distributed across all countries, that is, to avoid a situation where there are one or two countries that are much stronger than others,” DeLang said.