As AI wearables from companies like Meta continue to draw criticism over what they can see, hear, and record about the people around the wearer, Apple has been laying out a very different approach, and its executives are not shy about drawing the contrast directly. In a new interview with TechRadar, Apple’s senior vice president of hardware engineering John Ternus and its chief marketing officer Greg “Joz” Joswiak addressed the privacy backlash head on, using Apple’s new audio intelligence features on Apple Watch as the example of how they believe this kind of technology should actually work.
The conversation started with a pointed question about what some critics have started calling “pervert glasses,” a reference to the backlash Meta and other companies have faced over smart glasses that can quietly capture audio and video of everyone nearby without their knowledge or consent. While Ternus would not comment on any unannounced Apple products, he was clear that privacy and security shape every decision the company makes when building something new, and that this thinking guided the design of Apple’s own audio intelligence features from day one.
Why Siri Recap Has No Recording and No Transcript
Ternus pointed to Siri Recap as the clearest example of that philosophy in action. He explained that the ability to record audio or generate a transcript using apps on a watch or phone has existed for a long time, but Apple deliberately built Siri Recap to work differently. There is no recording created at any point, and there is no transcript generated either, not even one that the owner of the Apple Watch could go back and view later. According to Ternus, the recording simply does not exist, and neither does any form of speaker attribution, meaning the feature cannot identify or label who said what during a conversation.
Joswiak went further, explaining the technical reasoning behind that decision. He said a lot of people assumed the new audio intelligence features meant Apple Watch would be recording users constantly throughout the day, but that is not how the feature works at all. To make that possible, Apple built an entirely new chip for Apple Watch that includes what Joswiak called an exclave, a dedicated, isolated space that holds audio only for the brief moment it takes to process and generate high-level notes about a conversation. Once that processing is complete, the raw audio itself is gone.
The Case for Building Nothing That Can Be Hacked or Handed Over
Joswiak framed this as more than just a design choice, calling it a direct security benefit. If Apple never creates a recording in the first place, then there is nothing for anyone else to access, whether through a hack, a legal demand, or any other route. He acknowledged that people clearly want and need a way to stay present during meetings and conversations without worrying about capturing every detail themselves, which is exactly the gap this feature is meant to fill. But he was direct in saying Apple wanted to solve that problem without creating the kind of privacy exposure that other products on the market are currently dealing with.
The distinction Apple is drawing here is deliberate, and it comes at a moment when trust in always-listening, always-recording devices is under real scrutiny. By building a feature that generates useful, high-level notes about a conversation while making sure no actual recording or transcript ever exists, Apple is positioning itself as taking a fundamentally different approach than companies racing to add cameras and microphones to glasses people wear on their faces all day. Whether that approach becomes the industry standard or simply Apple’s own selling point remains to be seen, but for now, it gives the company a clear answer to one of the most uncomfortable questions facing the wearable AI category.
