Apple Confirms Siri's New AI Features in China Run on Alibaba's Qwen Models, Not Gemini
Apple has confirmed that its new Siri AI features in China will be powered by Alibaba's Qwen models rather than Google Gemini, which handles Siri AI elsewhere. A published guide indicates users will need separate Qwen accounts, suggesting the privacy protections built into Apple's Private Cloud Compute won't apply.
Apple has confirmed that Siri's new AI features in China will run on Alibaba's Qwen models rather than Google Gemini, which powers Siri AI in most other markets. The disclosure, reported by Reuters and first spotted in an updated Apple support guide, raises questions about whether Chinese users will receive the same privacy protections as users elsewhere.
In most countries, Apple's Siri AI features are powered by Google's Gemini models, running either on Apple's own Private Cloud Compute (PCC) infrastructure or on dedicated Google servers built to PCC-equivalent privacy standards. Under that arrangement, Apple states that neither it nor Google can access user interactions with Siri.
China does not permit this setup. Chinese law requires that AI services offered in the country use domestically approved models, a rule widely understood to give the government visibility into AI usage that foreign-hosted, encrypted systems would prevent. Apple has therefore partnered with Alibaba, integrating the company's Qwen models into Siri and Apple's Writing Tools for Chinese users.
According to Apple's updated Chinese-language support guide, users who opt in can use Qwen through Siri for more detailed responses to certain requests, including analysis of photos and documents. Writing Tools will also be able to call on Qwen to generate text or images from a text description. Reuters reports the guide is specific to Mac, but the same Qwen integration is expected to extend to iPhone, iPad, and Vision Pro in China.
A detail in the guide points to a meaningfully different privacy model: Mac users must have their own separate Qwen account and be signed into it to use the feature. This stands in contrast to the Gemini-powered version elsewhere, where Apple states user interactions remain inaccessible to both Apple and Google. Requiring a standalone third-party account suggests Alibaba, and potentially Chinese regulators, may have visibility into queries that Apple's PCC architecture is specifically designed to prevent in other markets.
Apple has not detailed what data protections, if any, apply to the Qwen integration, and the company has not commented beyond the support guide itself. Alibaba has not issued a statement on the arrangement's privacy terms.
What this means
This is not a new model release — it's a regional deployment decision forced by Chinese law, which bars foreign AI models from operating domestically. Apple's dependence on Alibaba's Qwen underscores a structural reality for any global tech company operating in China: local infrastructure and licensing requirements often mean privacy architecture cannot be uniform across markets. The requirement for a separate Qwen account is the clearest signal that Apple's zero-access privacy guarantee, central to its marketing of PCC and Gemini-powered Siri elsewhere, does not extend to Chinese users. Expect scrutiny from privacy researchers and possibly regulators outside China over whether Apple's global privacy claims about Siri AI need clearer country-specific caveats.
Related Articles
GitHub Models Retired, Cutting Off Free LLM API Access for GitHub Actions
GitHub Models, the unified LLM API and playground that let GitHub Actions workflows make free AI calls using existing GitHub credentials, has been fully retired. GitHub has not disclosed a reason, though rising inference costs from coding-agent usage are the likely driver.
Anthropic Makes Claude Code's Auto Mode Default for Pro, Max, and Team Users on August 14
Anthropic will make Claude Code's auto mode the default for Pro, Max, and Team accounts starting August 14, reducing step-by-step approval prompts. The company cites a study of 1,053 testers showing auto mode caught 89% of harmful actions versus 13.6% for manual review.
Anthropic Makes Auto Mode Default in Claude Code for Pro, Max, and Team Plans Starting August 14
Anthropic will make auto mode the default setting for new Claude Code sessions on Pro, Max, and Team plans starting August 14, 2026. The company cites a 1,053-person study showing auto mode blocked 89% of harmful actions compared to 13.6% for human reviewers, plus a third-party test claiming zero successful prompt injections out of 720 attempts.
Anthropic Makes Claude Code's Auto Mode Default, Cutting Dangerous Command Approvals by 75 Percentage Points
Anthropic will enable Auto Mode by default in Claude Code for Pro, Max, and Team plans starting August 14, 2026. Internal testing with 1,053 paid users found the classifier-based system caught 89% of dangerous commands compared to 13.6% caught by human reviewers, while producing 25% more pull requests.
Comments
Loading...