Apple accidentally revealed one of the clearest hints yet about its future plans for AirPods when a video demo of unannounced earbuds with built-in cameras was found in a macOS update.

The video, which appeared in a preview build of macOS Tahoe 26.7 before any official product announcement, suggests that Apple is testing the integration of vision and artificial intelligence technologies directly into its wireless earbuds.

The idea, as shown in the video, isn't to turn AirPods into a traditional camera for taking photos and videos, but rather to use optical sensors to give the earbuds the ability to understand the user's surroundings and connect what they see to artificial intelligence services and the Siri voice assistant.

What did the video show? The clip shows a person wearing AirPods holding up a book in front of them, allowing the earbuds to recognize what they are looking at through the built-in camera. After recognizing the book, the user can speak to Siri and ask it to be saved for later.

This demonstration offers a simple glimpse into how the earbuds could work in the future: instead of taking out an iPhone, opening the camera, and pointing it at something, the user could look directly at it and ask a voice question about it.

In other words, AirPods could become an additional eye for Siri, allowing the assistant to understand what the user is saying without needing to describe everything verbally.

What is "visual intelligence"? The idea relies on Visual Intelligence (VI), a technology Apple already uses in its devices to recognize objects, text, and information displayed in front of the camera. For example, the technology allows users to point the camera at an object, location, or text and retrieve related information, or perform actions based on what's visible. However, bringing this technology to AirPods represents a significant shift in how it's used. Currently,

the phone requires the user to take it out and point the camera manually, while the AirPods are wearable devices that can be worn for extended periods. If they can "see," AI can gain a direct visual context of what's happening in front of the user.