Jakub Porzycki Nurphoto Getty Images
MicrosoftOpenAI-backed ChatGPT has prevented users from using it in China, but the app has attracted huge interest in the country, with companies rushing to integrate the technology into their products and launch competing solutions.
While residents of these countries cannot create OpenAI accounts to access artificial intelligence (AI)-powered chatbots, virtual private networks and foreign phone numbers help overcome some of these restrictions.
At the same time, the OpenAI model behind the ChatGPT program, which can write essays, recipes and complex computer code, is relatively accessible in China and is increasingly integrated into Chinese consumer technology applications from social networking to online shopping.
The tool’s popularity has quickly raised awareness in China of US AI progress and, analysts say, just how far tech companies in the world’s second-largest economy have come as they struggle to catch up.
“There is great excitement around ChatGPT. Unlike the metaverse that faces great difficulties in finding real-life applications, ChatGPT has suddenly helped us achieve human-computer interaction,” said Ding Daoshi, director of Beijing-based internet consultancy Sootoo. “The changes we’re going to make are faster, more direct and faster.”
OpenAI or ChatGPT itself is not blocked by Chinese authorities but OpenAI does not allow users in mainland China, Hong Kong, Iran, Russia and parts of Africa to register.
OpenAI told Reuters it is working to make the service available.
“While we want to make our technology available everywhere, conditions in certain countries make it difficult or impossible to do so in a way that aligns with our mission,” the San Francisco-based company said in an emailed statement. “We are currently working to increase the number of locations where we can provide safe and meaningful access to our devices.”
In December, Tencent Holdings’ WeChat, China’s largest messaging app, has shut down some ChatGPT-related programs that appeared on the network, according to local media reports, but it continues to grow.
Dozens of bots rigged for ChatGPT technology have appeared on WeChat, with hobbyists using them to create automated programs or accounts that can interact with users. At least one account charged users a fee of 9.99 yuan ($1.47) to ask 20 questions.
Tencent did not respond to Reuters’ request for comment.
ChatGPT supports Chinese language interaction and can chat in Chinese, which helps drive informal adoption in the country.
Chinese companies are also using proxy tools or existing partnerships with Microsoft, which has invested billions of dollars in OpenAI, to access tools that allow them to embed AI technology into their products.
Proximai in Shenzhen in December introduced virtual characters to a 3D game-like social application that uses ChatGPT’s basic technology to chat.
Beijing-based entertainment software company Kunlun Tech plans to integrate ChatGPT into the Opera web browser.
SleekFlow, a Tiger Global-backed startup in Hong Kong, says it is integrating AI into its customer relationship messaging tools.
“We have clients all over the world,” said Henson Tsai, founder of SleekFlow. “Among other things, ChatGPT performs excellent translations, sometimes better than other solutions available on the market.”
censorship
Reuters tests of ChatGPT show that the chatbot does not reject sensitive questions in mainland China. Asked about his thoughts on Chinese President Xi Jinping, for example, he responded that he had no personal opinion and offered a variety of views.
But some proxy bots on WeChat have blacklisted the term, according to another Reuters investigation, in compliance with China’s heavy censorship of cyberspace. When asked the same question about Xi on one ChatGPT proxy bot, he responded by saying that the chat was against the rules.
To comply with Chinese regulations, Proximai founder Will Duan said the platform will filter the information provided to users during interactions with ChatGPT.
The Chinese regulator, which last year introduced rules to strengthen the governance of “deepfake” technology, has not commented on ChatGPT, however, the country’s media this week warned about the risk of the stock market amid the frenzy through ChatGPT shares-local concept.
China’s Cyberspace Administration, the internet regulator, did not respond to Reuters’ request for comment.
“With the regulations launched last year, the Chinese government said: we have seen this technology coming and we want to go further,” said Rogier Creemers, assistant professor at Leiden University. “I expect the majority of AI-generated content to be non-political.”
Chinese rival
Adding to the buzz have been some of the country’s biggest tech giants such as Baidu and Ali Baba who gave an update this week on the AI models they’ve been working on, so their stakes have been raised.
Baidu said this week it will complete internal testing of “Ernie Bot” in March, a large AI model the search company has been working on since 2019.
On Wednesday, Alibaba said its research institute Damo Academy was also testing a ChatGPT-style tool.
Duan, whose company has used Baidu’s AI chatbot named Plato for natural language processing, said ChatGPT is at least one generation more powerful than China’s current NLP solution, although it is weaker in some areas, such as understanding the context of the conversation.
Baidu did not respond to Reuters’ request for comment.
Access to OpenAI’s GPT-3, or Generative Pre-trained Transformer, first launched in 2020, an update that underpins ChatGPT.
Duan said long-term compliance risks mean Chinese companies will replace ChatGPT with local alternatives, if they can match the functionality of US-developed products.
“So we’re really hoping that there’s an alternative solution in China that we can use right away … maybe we can handle the Chinese people better, and we can also be more compliant with the regulations,” he said.