Apple announced long-anticipated AI upgrades to its Siri voice assistant this week at its annual World Wide Developers’ Conference. The new features are set for public launch later this year, but will not initially be available to users in the E.U. (except on Macs and Vision Pro headsets) or in mainland China due to regulatory issues.
For Apple users in China, the resulting situation is what the following meme describes as 一[苹]果两制 yī [píng]guǒ liǎng zhì, or "One Apple, Two Systems"—a pun on the official formulation 一国两制 yī guó liǎng zhì, or "One Country, Two Systems," referring to the principle by which Hong Kong and Macau exercise increasingly limited autonomy within China.
“Others are already using AI. You’re still waiting for Siri to evolve
One world, different permissions
Apple Intelligence
iOS 27 • International Edition
‘One Apple, Two Systems’
One world, two kinds of Siri.”
In the E.U.’s case, the issue is the requirement under the Digital Markets Act that third-party AI services must receive the same data and system access as Apple’s own. The company issued a lengthy statement this week arguing that this would present unacceptable privacy and security risks, and criticizing European regulators’ "extreme interpretation" of the law and refusal to accept compromises it had offered.
Curiously, Apple does not appear to have published any similarly in-depth explanation and objection in China’s case, offering only a footnote saying that "Siri AI and the other new Apple Intelligence features will not be available in China while Apple works through regulatory requirements." The company has worked with local partners to enable a Chinese launch of earlier Apple Intelligence features, which were announced in mid-2024, but these efforts have yet to bear fruit. Apple Intelligence features did appear briefly on China-based devices in March of this year, but this was reportedly an error: Bloomberg Apple reporter Mark Gurman posted on X that "it’s been ready to go for months but Apple doesn’t yet have regulatory approval."
Political guardrails will be key among the regulatory requirements in question. The Wall Street Journal’s Stu Woo noted in December that domestic chatbots must undergo internal and external safety checks related to 31 risk categories, with “incitement to subvert state power and overthrow the socialist system” topping the list. (Other risks cover less contentious ground such as promotion of violence or discrimination, or unauthorized use of others’ likeness.) One recent result is ByteDance chatbot Doubao’s refusal to explain the meaning of the English phrase "It’s my duty," which is associated with the 1989 Tiananmen protests.
Chinese users are hardly without domestic alternatives, and Apple is widely seen as lagging behind the competition even in markets where its AI features are actually available. But as illustrated by the controversial "Apple People vs Android People" discourse led by deplatformed former influencer Hu Chenfeng, the company and its products still have an aura of prestige and sophistication for many Chinese consumers. AI has also been more warmly embraced in China than in the West, so continuing to miss out on Apple Intelligence is galling for some.
For more on AI in the context of China’s online censorship, see China Media Project’s China Chatbot series, and related discussion in our interview with "Locknet" report authors Jessica Battke and Laura Edelson.
