Apple Intelligence meets Vision Pro
- A social post shared details suggesting Apple is integrating on‑device AI and eye‑tracking into Siri for spatial computing on Vision Pro. - The post framed this as part of an Apple Intelligence push toward more local model capabilities. - If accurate, that signals Apple leaning on on‑device features to differentiate spatial interfaces and privacy posture versus cloud‑first AI (x.com).
Apple has already put Apple Intelligence on Vision Pro, and the next question is whether Siri will start using the headset’s gaze controls more directly. (apple.com) Apple said on March 31, 2025 that visionOS 2.4 brought the first Apple Intelligence features to Vision Pro, including Writing Tools, Image Playground, Genmoji, smart replies, and ChatGPT integration. Apple’s support pages now list Apple Intelligence as built into visionOS and managed in Settings under “Apple Intelligence & Siri.” (apple.com) (support.apple.com) Vision Pro already lets people control interfaces by looking at an item and pinching, and Apple’s Siri guide says users can “open and interact with things you’re looking at” by voice. Apple’s developer documentation also shows visionOS is built around spatial sensors including hand tracking, world tracking, and room tracking. (support.apple.com) (developer.apple.com) That makes the reported Siri change plausible: in a headset, eye tracking works like a cursor, and an assistant that understands what you are looking at would have more context than a voice assistant on a phone. Apple has not publicly announced a new Siri feature for Vision Pro that matches the social post’s specific claim. (support.apple.com) (developer.apple.com) Apple has spent the past two years framing Apple Intelligence around local processing first. Its developer site says the system puts generative models “at the core” of iPhone, iPad, Mac, Apple Vision Pro, and Apple Watch, while Apple’s privacy materials say many requests run on device and more complex ones can go to Private Cloud Compute on Apple silicon servers. (developer.apple.com) (apple.com) Apple made that split more explicit at Worldwide Developers Conference in June 2025, when it said developers would get access to the on-device foundation model behind Apple Intelligence. Apple also said some features, including Live Translation, run entirely on device so conversations stay private. (apple.com) Siri’s current setup on Vision Pro is still mixed. Apple says voice input is processed on the headset, but transcripts of requests are sent to Apple to improve Siri, and some requests still require an internet connection. (support.apple.com) So the story here is less that Vision Pro suddenly got artificial intelligence, and more that Apple appears to be pushing the same formula deeper into spatial computing: local models where possible, cloud systems when needed, and interface cues drawn from where your eyes are already pointing. Apple’s public documentation already supports the first two pieces; the eye-tracking Siri claim still awaits an official product announcement. (apple.com 1) (apple.com 2) (support.apple.com)
