The Biden administration intends to impose restrictions on US-developed artificial intelligence (AI) models, such as ChatGPT, which are used in popular chatbots, Reuters reported.
The aim is to safeguard this technology from countries like China and Russia. However, China has been making strides in its domestic generative AI industry over the past year and has been urging its companies to steer clear of foreign technology.
Accessibility of OpenAI’s AI Models in China
OpenAI’s primary AI services, like ChatGPT and DALL-E, have not been formally launched in mainland China due to undisclosed “local conditions,” according to an OpenAI spokesperson quoted by Reuters. Nonetheless, many companies and engineers in China have managed to access OpenAI’s services using proxy tools like virtual private networks (VPN) to conceal their network addresses. Consequently, numerous Chinese companies have developed software and applications based on OpenAI’s models, often using them for benchmarking purposes.
Although OpenAI has taken steps to restrict access to its services by Chinese companies, including suspending the account of ByteDance, the Chinese owner of TikTok, Chinese entities continue to find ways to leverage AI technology developed by Western companies. In Hong Kong, access to OpenAI’s AI models is also limited, but not entirely blocked. Despite OpenAI’s services being unavailable in the region, Microsoft, an investor and partner of OpenAI, has released Copilot, a generative AI service based on OpenAI’s latest technology, to the public. Through collaboration with Microsoft, companies in Hong Kong can gain access to OpenAI’s AI models.
Utilization of US Technology in China’s AI Models
While the US Commerce Department’s focus is primarily on controlling the export of proprietary AI models, China’s reliance on open-source models developed in the West remains significant. Chinese entities, including prestigious research labs and high-profile AI companies like 01.AI, have been found to use open-source AI models from companies such as Meta Platforms (formerly Facebook). Despite these dependencies, several Chinese tech companies, including Baidu, are actively working to develop proprietary AI models to reduce reliance on foreign technology.
Beijing’s Perspective on US AI Models
In line with Chinese President Xi Jinping’s directive to achieve technological self-sufficiency, Chinese authorities have emphasized the importance of developing “controllable” AI technology domestically. State-backed media outlets like China Daily have raised concerns about the potential risks associated with US generative AI technology, suggesting that it could be used for spreading disinformation and manipulating global narratives for geopolitical interests. The Chinese government has also implemented regulations on the use of generative AI, requiring government approvals for services before they can be released to the public.
Positive sentiments from the Chinese government towards US generative AI technology have primarily been framed in terms of acknowledging the gap between China and the US in AI development, rather than actively promoting US AI technology. Analogies like comparing ChatGPT’s superiority to Chinese AI products to Messi’s football prowess highlight the challenges China faces in catching up with the US in AI development.
Source: reuters.com
