Illinois Photos.app People Album Lawsuit
Apple faces a $32.5 billion lawsuit over how its Photos app scans and stores faces. A federal judge cleared the case to proceed as a class action late last month, allowing millions of Illinois iPhone owners to join.
The case has been years in the making, and Apple has fought to shut it down at nearly every turn. Now it heads back to the court, with the company’s privacy practices on trial.
[…]
Court records have shown that the update applies to users with iCloud Photos turned on. Those accounts also need to use at least 10GB of storage and contain at least 5,000 photos and videos. Plaintiffs say all this amounts to collecting biometric data without consent, putting Apple squarely at odds with Illinois law.
Again, I find it odd that features are apparently gated based on the capacity of your iCloud account rather than the available storage.
The lawsuit accuses Apple of collecting biometric data without user consent in the Photos app, reports The Times. It is alleged that Photos uses facial recognition technology to scan individuals who appear in images, creating a “faceprint” for each person in the photo library.
After collecting enough samples, an algorithm is allegedly used to identify the iPhone user. That data is then stored on the iPhone within the Photos app.
I don’t see what the issue is here. My understanding is that it only uploads vectors after you’ve opted in by enabling iCloud Photos and manually assigning names. Also, the vectors don’t seem like the sort of biometric information that the law was meant to protect.
Assuming things are working properly, what should happen is that the people you have specifically named and the photos you have specifically verified sync their info via iCloud. However, the actual face detection for the rest of the photos is done on a per device basis and is not synced via iCloud, primarily for security/privacy reasons.
So, when you first set up a new device, it will pull down the names and info for the photos where you have identified a particular person, but then it still needs to go through the entire library on-device, identify the faces in each photo, and compare them to the synced information to tell which person is in which photo. So once it’s done, you should end up with mostly the same photos identified as each person, but each device still needs to do its own face detection locally.
