Apple is preparing what sources describe as its most significant Siri overhaul in fifteen years, set to debut at WWDC in June. According to Bloomberg's Mark Gurman, the centerpiece of the revamp is a standalone Siri app—the first time the assistant has existed as something you open rather than invoke—featuring an experience described as reminiscent of ChatGPT. But the headline feature might be one that actually respects user data: an auto-delete option for conversations modeled after how Messages handles retention, letting users set chats to expire after 30 days, a year, or keep them indefinitely.

What Apple Is Actually Building

The new Siri will ship with meaningful privacy controls built in from launch. Users get explicit options around conversation retention—something most AI chatbots don't offer at all. Apple's framing centers on storing less, retaining less, and giving users more control than any competitor. For a company that has cultivated serious brand equity around privacy, leaning into this for an AI relaunch makes obvious strategic sense. The standalone app format is also significant; it represents a fundamental shift in how Siri exists within iOS rather than being woven into system-level functions.

The Gemini Problem Nobody's Talking About

Here's where Apple's narrative gets complicated. The model doing the actual intelligence work inside this new Siri isn't Apple's. It's Google Gemini, under a deal reportedly costing Apple roughly $1 billion annually. When you ask the new Siri something that requires real reasoning, your query routes to Google's infrastructure. Apple can control its own server-side retention policies and build auto-delete into the app interface, but what happens once that request hits Google's servers is largely out of Apple's hands—a tension Gurman flagged directly in his reporting.

Privacy Pitch or Convenient Cover?

There's an uncomfortable possibility here that WWDC will either confirm or deny. A privacy-first framing genuinely differentiates Apple from competitors who treat user conversations as training data. It's also a remarkably convenient explanation for why the product might do less than expected, because privacy constraints wouldn't allow it. The gap between "Apple takes privacy seriously" and "Apple routes your queries through Google" is noticeable. Whether Apple's executives address this directly on stage or hope developers and press connect dots the average iPhone user won't bother to trace remains one of the more interesting questions about next month's presentations.

What WWDC Will Actually Answer

June is close enough that this is less speculation and more preview. The broad strokes—a standalone app, chatbot experience, robust privacy controls, Gemini backend—are consistent with everything reported about Apple's AI direction over the past several months. The auto-delete feature deserves credit regardless of the backend question; user control over conversation retention should be standard across all AI products and most still don't offer it. But WWDC will reveal whether Apple acknowledges the Google connection to general audiences or treats it as something people need to already know to understand.

Key Takeaways

  • Standalone Siri app with ChatGPT-like experience coming at WWDC 2026
  • Auto-delete conversations (30 days, 1 year, indefinite) gives users real retention control
  • Privacy pitch is genuine differentiator but also convenient product limitation explanation
  • Google Gemini backend means Apple doesn't control the full AI pipeline despite privacy branding

The Bottom Line

Apple's privacy story only holds if you ignore where the actual intelligence happens. Auto-delete on Apple's servers matters, but when your queries are being processed by Google's infrastructure, you're trusting two companies with different incentives and different data practices. WWDC will show whether Apple is serious about transparency or just weaponizing privacy as a marketing shield.