Apple has a Siri problem and everyone knows it. ChatGPT became a verb. Gemini is powering half the Android ecosystem. Claude is showing up in enterprise workflows. Meanwhile Siri is still struggling to set timers reliably.
WWDC is in June and Apple is reportedly planning its biggest Siri overhaul yet. A standalone app, a proper chatbot experience, and a privacy pitch front and center. According to Bloomberg’s Mark Gurman, Apple executives plan to argue they’re taking a more privacy-friendly approach than every other AI company out there.
That argument gets complicated quickly. The model powering this new Siri is Google Gemini.
What Apple is actually building
The revamp has a few concrete pieces according to Gurman. First, a standalone Siri app, the first time Siri has existed as something you open rather than something you invoke. The experience is described as reminiscent of ChatGPT, which is a significant shift from what Siri has been for the last fifteen years.
One of the more specific details Gurman surfaced is an auto-delete option for conversations, similar to how the Messages app handles retention. Users would be able to set chats to delete after 30 days, after a year, or keep them indefinitely. That kind of user control over data retention is meaningful and more explicit than what most AI chatbots currently offer.
Apple’s framing will reportedly center on the argument that it stores less, retains less, and gives users more control than competitors. For a company that has built significant brand equity around privacy, leaning into that for its AI relaunch makes obvious sense.
You May Like: OpenAI Wanted Distribution on the iPhone. Apple Had Other Plans.
The Google problem
The model doing the actual work inside this new Siri is Google Gemini, under the deal Apple struck earlier this year paying roughly $1 billion annually. When you ask the new Siri something that requires real intelligence, that query is going to Google’s infrastructure.
Apple can control the retention policy on its end. It can build auto-delete into the app. It can limit how long conversation data lives on Apple’s servers. What it has significantly less control over is what happens on Google’s end once the request gets there.
Gurman flagged this tension directly, suggesting Apple’s privacy emphasis might also be functioning as cover for Siri’s limitations compared to ChatGPT and other dedicated AI products. A privacy-first framing is genuinely differentiating. It’s also a convenient way to explain why the product does less because privacy constraints wouldn’t allow it.
Whether that’s the full part or a convenient one is something WWDC will at least partially answer.
What to actually expect at WWDC
June is close enough that this is less speculation and more preview. The broad strokes Gurman is describing, standalone app, chatbot experience, privacy controls, Gemini backend are consistent with everything that’s been reported about Apple’s AI direction over the past several months.
The more interesting question WWDC will answer is how Apple handles the Google disclosure. Right now the privacy pitch and the Gemini backend exist in separate conversations. On stage in June they’ll have to coexist in the same product demo. How Apple explains that to a general audience, not to journalists who already know about the deal, but to the average iPhone user who assumes Siri is an Apple product end to end, will say a lot about how honest this privacy relaunch actually is.
The auto-delete feature is useful regardless of the backend question. User control over conversation retention is something every AI chatbot should offer and most don’t. Apple building that in from launch is worth acknowledging.
But the gap between “Apple is taking privacy seriously” and “Apple is routing your queries through Google” is a noticable gap. June is when Apple has to close it or explain why it doesn’t need to be closed.
We’ll be watching.
