The system no longer provides FB the same value it previously did. So they're taking it down, now that they've collected billions of images. Will they be deleting the trained models too? All artifacts and inferences? Hell no.

"the social network will delete the face scan data of more than one billion users."

I believe it because facebook has nothing much to gain by lying about this small thing(after all they still have all the tracking data) and hell lot to loose by lying this straight face among employees.

> and hell lot to loose by lying this straight face among employees.

And if, in a year an employee speaks out and leaks that FB never deleted the data? What would happen then? I don't think anything would happen, they wouldn't lose much, and no one would be too surprised.

Normally the cost of cognitive dissonance in tech is $200k/yr. Pay that and you will have a line out your door and around the block of engineers who want to work for you, no matter what you do.

Fortunately Facebook pays for the Cognitive Dissonance Deluxe package which is about $250k-$400k year. They've got plenty of padding to mislead their employees right to their face.

“We didn’t realize that our old archive data was retained. Whoopsie.”

Not. It does _not_ say "all users". Additionally, there's a lot of weasel words.

It does not say they've stopped doing facial recognition. It just says they've shut down one program.

They also specifically reference the "usage" of facial recognition, not the act of facial recognition.

My guess is they are still being paid by governments to collect this data but are just not exposing it.

They have then politicized it by referring to "regulators" and "regulations" several times.

fb just had to pay $650m to people in Illinois for blatantly violating their biometrics law.

I suspect that between Illinois and the EU, there's flatly no way to reliably legally do this.

Yeah, I could see Illinois saying that scanning a picture for only opt-in templates is still "scanning" people who haven't opted in because it does face detection and FB not wanting to take it to the Supreme Court.

To me FB is the embodiment of that old joke about politicians: You know they're lying because their mouth is moving.

I do wonder if they are merely substituting outright facial recognition with some other technology where they are able to identify a person in a particular piece of content. It could be text associated with content or the geographical proximity of people at the time content was captured or published.

> Trying to figure out their real angle here.

Maybe that most phone cameras are now good enough to enable iris scanning, which is orders of magnitude more accurate.

They want the feature regulated/pariahed so no other social network can leverage it.