Menu Menu

The ChatGPT black market is thriving in China

China clamped down on access to ChatGPT last month, announcing its own generative AI chatbot ‘ERNIE Bot,’ arriving in March. This has since ignited an illicit marketplace for OpenAI’s hot property software within China.

The commercial road-map for novel chatbots has been established: complete integration into their parent company’s internet search engines.

Microsoft-owned OpenAI has already combined ChatGPT with Bing, setting an early benchmark for quality within generative text AI. Snapchat has since announced an AI feature of its own, and Alphabet is scrambling to prepare Bard for Google Search after a bumpy start.

Outside of the west, China quickly followed Google by unveiling an upcoming chatbot ‘ERNIE Bot’ – or Enhanced Representation through Knowledge Integration – slated to merge with its main search engine Baidu in March. The same day, Baidu’s shares surged 15% on the Hong Kong stock exchange.

There are several other iterations of the technology being developed in China, including HunyuanAide and IFlyTech, but crucially, avenues to ChatGPT have been blocked by the nation’s online regulators due to its inherent US links.


China’s call of ‘political propaganda’

Talk of innovation with AI is as feverous as you’d expect in China, but tech companies are often at odds with state government and progress is being tempered.

With increasingly strict internet censorship rules already in effect, China is unwilling to open the floodgates to just any new product, taking particular exception to software developed by countries deemed as political adversaries.

Its tech firms and universities are pushing to fill the domestic void of AI bots, but in the meantime a full-scale ban on ChatGPT has been put in place.

In February, an official statement revealed ChatGPT had been indefinitely suspended on the grounds that ‘the model is trained on information based in western countries,’ and that its answers on Chinese politics were ‘consistent with the political propaganda of the US government.’

Despite these regulations, with no chatbot alternatives to use for recreational or work purposes, Chinese users have found sneaky ways of circumventing censorship and logging back on to ChatGPT.


The illicit marketplace for ChatGPT

ChatGPT logins have become a hot commodity on Chinese shopping platform Taobao, akin to that of foreign phone numbers able to receive verification codes.

A search in early February returned more than 600 vendors for ChatGPT logins, some with thousands of individual sales at prices ranging from the equivalent of $0.17 to $4.28. In the majority of cases, intermediaries were involved, meaning someone with access to ChatGPT would ask a question on someone’s behalf for a fee.

On the aforementioned Baidu, somewhat ironically, the phrase ‘How to use ChatGPT within China’ had trended for weeks and ultimately alerted Chinese authorities to the scope of the problem.

WeChat parent Tencent and Ant Group has been ordered to cut access entirely in late February, but VPNs and third-party developers will no doubt persist in the months ahead. Whether this continues beyond the rollout of domestic AI chatbots remains to be seen.

Given the strict precedent set by online authorities in China, it’s not unlikely that ChatGPT will live on in some capacity. China’s social media platforms employ barriers of algorithms and human moderation to censor sensitive information, and chatbots will not be exempt.

‘It is likely that we will see this type of human-reliant censorship, in combination with other tactics like keywords blocking, being used in public-facing chatbots,’ says Thomas Quitong Cao, a political behaviour expert on the internet.

There’s something that feels intrinsically wrong about supposedly ‘limitless’ technology being stifled by stringent regulations, and this may plague China’s generative AI endeavours for the remainder of the year and beyond.

Accessibility