In the wave of generative AI reshaping human-computer interaction, Apple is preparing a key counterattack over "the authority of smart assistants." According to media reports, Apple will launch a new version of Siri at its Worldwide Developers Conference (WWDC) in June. Privacy protection will be the core selling point — including the ability for users to set automatic deletion of chat records (after 30 days, one year, or permanently). This design clearly targets competitors like ChatGPT, aiming to establish a differentiated advantage through a "more restrained data strategy."

Privacy First: A Genuine Commitment or a Strategic Rhetoric?

The new Siri is rumored to debut as Apple's first standalone app and introduce a conversational experience similar to ChatGPT. However, unlike most competitors, Apple emphasizes:

  • Data retention control: Users can choose how long to keep conversations, with automatic deletion after the set period;
  • On-device processing priority: As much reasoning as possible is done locally on the device, minimizing cloud transmission;
  • Minimal data collection: Only necessary contextual information is collected to complete tasks.

This strategy aligns with Apple’s longstanding "privacy as a human right" brand narrative and responds to global regulatory trends such as the EU Digital Markets Act and China's Personal Information Protection Law. However, industry observers point out that emphasizing privacy may also become a "cover-up": when Siri still lags behind competitors in hard metrics like task completion, multi-turn reasoning, and tool invocation, the rhetoric of "we protect you more" might ease user doubts about product capabilities.

Secret Collaboration: Is Gemini Powering Siri?

More intriguingly, Gurman revealed that the new Siri may partially rely on Google's Gemini model. If true, this would create a delicate situation:

  • Front-end: Apple presents itself as a "privacy guardian" to users;
  • Back-end: Google provides large model inference capabilities, handling some cloud requests.

This "Apple-designed + Google-powered" hybrid architecture allows Apple to quickly fill technical gaps without fully relying on third parties, which could harm brand independence. But it also raises new questions: When users believe their data only flows within the Apple ecosystem, it may actually be processed through Google servers — where exactly does the boundary of privacy commitments lie?

Independent App: Siri's "Second Curve"?

Releasing an independent Siri app is seen as a key step for Apple to break free from the limitations of a "system-level assistant":

  • Expanded scenarios: No longer limited to "Hey Siri," users can open the app directly for deep conversations and task planning;
  • Enhanced experience: Supports rich text and multimodal input, closer to a professional AI assistant;
  • Ecological synergy: Deep integration with Apple Intelligence, iCloud, and HomeKit, building a "personal intelligent hub."

But challenges remain significant: if the standalone app still frequently calls cloud-based large models, the on-device privacy advantages will be diluted; if features are overly restricted to ensure privacy, practicality may be sacrificed. Finding a balance between "security" and "intelligence" will be key to regaining user trust.

🌍 Industry Insight: Can Privacy Become the New Moat in AI Competition?

Apple's attempt raises a new question for the global AI industry: As large model capabilities gradually converge, could data governance strategies become the next focal point of competition?

  • For developers: More precise trade-offs are needed between model performance and privacy compliance;
  • For users: They will have more "voting with their feet" power, choosing AI products aligned with their values;
  • For regulators: Clearer rules are needed regarding cross-border data flows and liability identification.