Pat Dolan has lived with ALS for a decade. What he wants, he says, is the option to steer his power wheelchair himself — and Apple says that capability is coming to its Vision Pro headset. The company on Tuesday previewed a broad set of accessibility updates powered by Apple Intelligence, including natural-language controls for VoiceOver and Voice Control, on-device subtitles for uncaptioned video, and eye-tracking control of compatible power wheelchairs through the headset. The features will arrive later this year as part of Apple’s next software cycle.
The announcement, ahead of Apple’s usual fall release window, pairs the company’s on-device AI with its longest-standing product priority. “Apple’s approach to accessibility is unlike any other,” Chief Executive Tim Cook said in the press release. Sarah Herrlinger, Apple’s senior director of global accessibility policy and initiatives, said the updates bring “new, intuitive options for input, exploration, and personalization — designed to protect users’ privacy at every step.”
VoiceOver, Apple’s screen reader for blind and low-vision users, gains an Image Explorer that uses Apple Intelligence to deliver detailed system-wide descriptions of visual content — personal photographs, scanned bills, receipts and documents. Live Recognition, meanwhile, now integrates with the Action button on supported iPhones: a user can point the camera at an object or scene, ask a question about what is in the viewfinder, and get a detailed audio response, with follow-up questions allowed for more specific context. Accessibility Reader gains the same generative description capabilities, and the updates are designed to work across the ecosystem rather than on a single device.
Magnifier benefits from similar generative tools and now supports spoken requests, so users can simply say “zoom in” or “turn on flashlight” rather than navigating menus. Voice Control gets the largest conceptual change: instead of forcing users to memorize exact button labels or grid numbers, the system now understands natural-language descriptions of onscreen elements — a user can say “tap the purple folder” and the system finds it. For users with limited dexterity or vision, that is the difference between navigating a phone by memorization and navigating it by description.
Apple is also bringing on-device generated subtitles to any video that lacks captions — personal recordings, shared clips and streamed content — across iPhone, iPad, Mac, Apple TV and Vision Pro. The system uses on-device speech recognition to transcribe spoken audio automatically, which means the processing stays on the device rather than in a cloud; there is no cloud upload and no hosting by Apple, consistent with the privacy framing the company attaches to its entire AI strategy. Subtitle appearance can be customized in the video playback menu or in Settings. For users with hearing loss, the feature closes a gap that has existed since the first smartphone videos: uncaptioned content from friends and family.
The most visible addition is the power wheelchair control feature for Vision Pro. Using the headset’s precision eye-tracking system, users can drive compatible power wheelchairs through alternative drive systems, with eye tracking that Apple says requires no frequent recalibration and works across a variety of lighting conditions. The feature launches in the United States with Tolt and LUCI alternative drive systems, supporting both Bluetooth and wired connections, and Apple says it will work with developers to expand support to more drive systems. “The option to control my power wheelchair on my own is gold to me,” said Dolan, founder of GeoALS and a member of Team Gleason’s patient advisory board. Blair Casey, Team Gleason’s chief executive, said the industry has seen eye-driven wheelchair systems evolve over the past decade and called Apple’s use of Vision Pro eye tracking “a huge step forward.”
The engineering claims matter for real-world use: eye tracking that required frequent recalibration would be impractical in a device meant to be used daily, and Apple says Vision Pro’s system holds calibration across the lighting conditions users encounter between indoor and outdoor settings.
The accessibility push doubles as Apple’s most tangible demonstration of Apple Intelligence outside text and image generation — AI that navigates the physical world for users, rather than producing content for them. The timing, just ahead of Google’s annual developer conference, also gives Apple a product story centered on on-device privacy at a moment when AI’s data practices are under scrutiny.
There is also a commercial subtext. Accessibility features have historically driven hardware adoption for Apple, and the Vision Pro — a product that has struggled to find its market — gains a use case with a concrete, emotional demonstration of what spatial computing can do. Alongside the software, Apple began selling the Hikawa Grip & Stand for iPhone, an adaptive MagSafe accessory designed with accessibility at its core, in three new colors starting Tuesday. VisionOS 27, the platform the wheelchair feature will ship on, also adds vehicle motion cues to reduce motion sickness, face gestures for system actions and new eye-selection controls using dwell time.
The features arriving later this year will test whether Apple Intelligence can deliver on accessibility — the product category where the company’s privacy stance and its hardware depth matter most. If the eye-tracking claims hold up outside the demo room, the wheelchair feature could become the clearest proof yet that spatial computing has a role beyond entertainment and productivity.
