The Chinese regime has begun targeting domestic tech providers that help users access foreign artificial intelligence models, marking a new phase in Beijing’s efforts to tighten its control over AI services.
In recent days, multiple China-based AI relay platforms that provide access to models such as ChatGPT and Claude have suspended or scaled back their AI offerings.
The closures follow reports that police questioned some platform operators. Screenshots circulating online show an administrator of one large AI relay platform informing users that he had been ordered by police in the eastern city of Hangzhou to shut down any services operating without approval.
“After receiving a call from Hangzhou police, I was summoned and disciplined,” the administrator wrote in a group chat. “Authorities have ordered the closure of AI relay platforms that lack the relevant operating licenses. The service will cease operations starting today, and users with active subscriptions will receive refunds.”
The platform later announced it had shut down.
OpenAI and Anthropic do not offer services in China, so businesses in China set up relay platforms allowing Chinese developers, businesses, and individual users to access models such as ChatGPT and Claude without registering overseas accounts.
The action follows a June warning from China’s Ministry of State Security that publicly criticized AI relay platforms, according to state media China Central Television.
Some operators lacked the proper qualifications, posed risks of cross-border data transfers, and threatened national security, the ministry said. It urged users to choose “secure and compliant” AI services.
Demand for Uncensored AI
Despite the demand for access to foreign AI models, Chinese regulators last year started requiring that generative AI products register with authorities. Authorities also exercise oversight of all training data, AI-generated content, and user information.
By the end of April, China’s Cyberspace Administration reported that 868 generative AI services had registered, while 530 applications and functions that rely on foreign AI models had completed separate government filings.
A current affairs commentator, who spoke to The Epoch Times on condition of anonymity out of fear of reprisal, said the ruling Chinese Communist Party (CCP) remains concerned that overseas AI models “have not undergone China’s registration, content review, and data supervision requirements.”
That’s also why Chinese people are seeking them out, a Chinese scholar, who also remained anonymous for safety reasons, told The Epoch Times.
“Foreign AI models can answer questions that Chinese AI systems often sidestep, allowing users to access information beyond the Great Firewall,” he said.
Impact on Businesses
The expanded policing is affecting more than casual users.
A Chinese network engineer told The Epoch Times that many software developers and small businesses rely on AI relay platforms to test applications, process text, and power their AI customer-service tools.
“These platforms aren’t just for chatting with AI,” he said. “When a major relay platform shuts down, it disrupts not only ordinary users but also small development teams whose businesses depend on overseas AI models.”
The engineer said the campaign appears at odds with Beijing’s stated goal of promoting AI development.
“As larger platforms disappear, users are likely to migrate to smaller, more discreet services,” he said. For business operations, “that will increase both the difficulty of access and the risks involved.”
The scholar said the result will likely be even fewer information sources and technology options for users in China.
