Apple is showing the public today what its "artificial intelligence" looks like - and the biggest news is exactly what Apple doesn't want to say out loud: the new Siri will run on Google's technology. Specifically, on the Gemini model. The company that for years sold its software as closed, proprietary and independent is now building its most important AI product propped up on someone else's engine.

The revamped Siri will, according to the announcements, finally do what ChatGPT and Claude have been doing for some time - conversation with context, multi-step tasks, deep integration into apps. A standalone Siri app is also expected, one that will compete directly with those same products, with an option to automatically delete conversations after 30 days, one year, or never.

The rest is a catalog of AI features Apple is adding belatedly: "visual intelligence" in the camera (again propped up on Google's image search), editing photos with plain speech, automatically removing objects from images, AI agents in the App Store that will book a table or edit a document for you. Apple Wallet gets bill splitting and digital passes.

The question few ask out loud at presentations like these: if your biggest selling point for years was that you build everything yourself and don't trust others with users' data - what does it mean when you sign a deal with Google precisely to power your voice assistant? The user gets a better Siri. But do they get a more private Siri, or just a more nicely packaged Google presence in their phone?