Apple has reportedly removed more than 100 generative artificial intelligence apps from the Chinese version of its App Store believed to be in violation of the country’s strict new regulations clamping down on “subversive” speech produced by ChatGPT-style chatbots. Those restrictions, which officially take effect August 15, demand that AI developers’ products “adhere to core socialist values” and prohibit content that questions state power. If the developers can’t control their generative progeny, the products will face censorship and possible shutdowns.
In a letter sent to developers shared on Twitter by native ChatGPT client OpenCat, Apple said it was removing the apps “because [they] include content that is illegal in China.” Many of the apps being removed appear to rely on OpenAI’s ChatGPT, which Apple said does not have a proper license or permit to operate in the country under the new guidelines. Popular Chinese AI apps like Spark, ChatGAi Plus, OpenCat, and ChatbotAI have all reportedly been removed as part of the government’s generative AI purge.
“As you may know, the government has been tightening regulations associated with deep synthesis technologies (DST) and generative AI services, including ChatGPT. DST must fulfill permitting requirements to operate in China, including securing a license from the Ministry of Industry and Information Technology (MIIT),” Apple said in its note to OpenCat. “Based on our review, your app is associated with ChatGPT, which does not have requisite permits to operate in China.”
Apple and OpenAI did not immediately respond to Gizmodo’s requests for comment.
China expands Great Firewall in the AI age
The removals mark the first stage of a new era of AI regulation by Chinese authorities which are simultaneously encouraging domestic companies to ramp up production of their own more tightly moderated large language models that could potentially rival OpenAI and Google. Last month, The Cyberspace Administration of China released new guidelines calling on AI developers to undergo security reviews and register their algorithms with government regulators. The guidelines prohibit companies from releasing AI models to the public that could subvert state power, compromise national security or “overthrow the socialist system.” Less controversially, the guidelines also mandate AI companies to take effective measures to prevent discrimination in or raining data and respect intellectual property rights.
The formal guidelines released in July actually eased up on several provisions first offered in an April draft proposal. Notably, the restrictions noted above won’t apply to research institutions or industries developing AI models that won’t be released to the wider public. This carve-out could give Chinese AI developers the flexibility to try and catch up to their US counterparts while still maintaining a tight grip over the content large language models eventually serve up to consumers.
Experts speaking with Gizmodo previously said these are part and parcel of broader internet restrictions in the country that suppress political speech in the name of security. Yaqiu Wang, a Senior China Researcher at Human Rights Watch, said Chinese regulators likely already possess the ability to punish AI companies for publishing supposedly subversive content but said these guidelines make that process much simpler and more streamlined. Others, like Article 19 Asia Digital Programme Manager Michael Caster, fear the new rules could have a wide impact and potentially cut off access to foreign articles translated by chatbots or suggestions on how to use VPNs to tip-toe around internet restrictions.
“We should be under no illusions. The Party will wield the new Generative AI Guidelines to carry out the same function of censorship, surveillance, and information manipulation it has sought to justify under other laws and regulations,” Caster said.