Google is expanding the kinds of user data it may save for AI training, including Lens photos, Search Live recordings, and Translate audio, according to The Verge. The change puts more multimodal interactions into the data-governance spotlight.
The concern is not only whether the data is useful for model improvement. It is whether users understand when images, voice, and translation sessions are retained and how to control those settings.
As AI moves into cameras, microphones, and live search, privacy choices become harder to notice and more important to get right.