Since then, though, the expectations of the level of Apple’s capitulation have only grown more onerous. Algorithms that determine what the public sees online or through AI have to be registered with the Chinese authorities, and new AI legislation is largely focused on controlling the exact public-interfacing models that Western tech companies want to get involved with.
“You need to file with regulators. You might need to submit a lot of details about things like coding … many tech companies may not be willing to do that,” says Tan.
The problem is, China can afford to put in place such measures because the power balance is in its favor—more so than ever.
“China is no longer just playing a following role in many technology fields,” adds Tan. “It is already advancing and taking the leading role.”
Business as Usual?
From a Western view, the rules put in place for generative AI in China veer between the admirable and the worrying.
“The regulation includes a number of vague censorship requirements, such as that deep synthesis content ‘adhere to the correct political direction,’ not ‘disturb economic and social order,’ and not be used to generate fake news,” reads Carnegie Endowment’s paper on the state of affairs in 2023.
“Deep synthesis” is the term the CAC uses in place of generative AI. China’s restrictions would result in a Siri that wouldn’t talk about the Dalai Llama, that wouldn’t refer to Taiwan as a separate country or acknowledge the Uyghurs. And who knows what else.
Given the current lax state of Western LLMs, it’s hard enough to picture a chatbot that couldn’t be cajoled into saying China is a part of the sovereign state of Taiwan, let alone falling into line 100 percent of the time. But clearly many Chinese tech companies have managed to adhere to the restrictions, to the satisfaction of the regulators at least. In August 2024, the South China Morning Post reported 188 LLMs had been approved for use to date, up from just 14 in January 2024.
It could be argued that Apple effectively adopting a custom version of one of these LLMs to fill out China’s version of Apple Intelligence represents business as usual. Apple already censors the app store to comply with China’s policies. It already cooperates with local entities.
However, with Apple Intelligence generative AI positioned at the heart of iPhones and other devices, the company seems more at risk of being accused of being a little too embedded in the wants and whims of the Chinese state for comfort, for a US company.
In August, Zhuang Rongwen, director of the Cyberspace Administration of China, said generative AI, such as chatbots, was “forcefully driving economic and societal growth.” The New York Times’ 2021 report suggested the government didn’t really need Chinese iPhone users’ data to surveil its citizens, as it already had stronger methods. But with GenAI, Apple may inadvertently become a more active participant in the CCP’s goals.