Facebook plans to shut down its facial recognition system | Hacker News Reader

Facebook plans to shut down its facial recognition system

nytimes.com

265 pointsnycdatasci186 comments

http://web.archive.org/web/20211102170829/https://www.nytime...

> The change affects more than a third of Facebook’s daily users who had facial recognition turned on for their accounts, according to the company.

I'm surprise the number is that low. Wasn't enabling this a default?

> That meant they received alerts when new photos or videos of them were uploaded to the social network. The feature had also been used to flag accounts that might be impersonating someone else...

That's actually a pretty valuable and legitimate use case for such a system, even from a very pro-privacy perspective.

> Although Facebook plans to delete more than one billion facial recognition templates, which are digital scans of facial features, by December, it will not eliminate the software that powers the system, which is an advanced algorithm called DeepFace. The company has also not ruled out incorporating facial recognition technology into future products, Mr. Grosse said.

It's also almost certainly not going to delete the face-specific image tagging either, which means they could re-derive everything they're "deleting" relatively easily.

>I'm surprise the number is that low. Wasn't enabling this a default?

The first implementation appeared in 2017 [1] as tag suggestions, but the facial recognition feature we know today evolved around 2019, when it was made opt-in rather than a default setting. [2]

[1] https://www.theverge.com/2017/12/19/16794660/facebook-facial...

[2] https://www.theverge.com/2019/9/3/20847650/facebook-facial-r...

Why does the NYT OP suggests Dec 2010?

> The decision shutters a feature that was introduced in December 2010 so that Facebook users could save time. The facial-recognition software automatically identified people who appeared in users’...

I actually have NYT blocked because their reporting has had such shoddy research lately, but I guess they're probably referring to the face-clustering feature for tags where, when you upload an album, they'll cluster all the like-faces to you to let you tag all of them at once [1].

It was released in December 2010 and, AFAIK, is technically only face detection (they didn't suggest whose face it might be for many years), not facial recognition. (They bought facial recognition company Face.com in 2012 and everyone was excited that FB might finally get real recognition).

I guess it could be argued that their implementation of face detection ultimately evolved into their current facial recognition systems (see: Ship of Theseus?), or that to a layperson the two technologies might (unfortunately) sound like the same thing. Or maybe I'm just being pedantic; I was just trying to add some historical context to the current, opt-in "facial recognition" feature in my comment above.

[1] https://www.theatlantic.com/technology/archive/2010/12/faceb...

No, you're not being pedantic, I appreciate the history, I was just wondering if anyone knew what was what here, after you supplied those citations for much later introduction date. And you supplied some info, thanks!

(I don't know why you have to block something for having bad reporting, but we don't have to talk about that!)

Thanks, glad I could assist!

(Haha, always trying to take back a few seconds/minutes in the day. I find blackholing links to a few websites that get linked often, but aren't often worth reading, helps with that. :) )

I'm still seeing things that suggest they really did deploy this in 2010 though... were they press-releasing more than they were actually doing in 2010?

June 2011...

> “When you or a friend upload new photos, we use face recognition software . . . to match your new photos to other photos you’re tagged in,’’ Justin Mitchell, a Facebook engineer, said in a Facebook blog. “We group similar photos together and, whenever possible, suggest the name of the friend in the photos.’’

http://archive.boston.com/business/technology/articles/2011/...

I turned it off because the "default" use case seemed rather to be to boost their engagement because someone you know socially would upload a picture of something they did that happened to have you in it. That would lead to you being tagged and getting notified, presumably to talk about how much fun you all had or whatever it was. Unfortunately it's pretty useless to ask people to not upload pictures that you're present in, and FB used notifications essentially just to get you to open the app and re-engage. I turned off everything that generated useless notifications for a while, then I finally turned off notifications completely.

Everything in FB is centered around the use case of boosting their engagement. That's pretty much the only metric they steer by.

The facial recognition feature is how I figured out someone was doxxing and harassing me on their page. Getting the notification that it thinks I appeared in a photo they had ripped from my FB profile was the tip off. I have mixed feelings.

But then, it's probably FB that facilitated that kind of behavior, so more mixed feelings.

Now there's one less tool to help the victim of such abuse though.

This comment to me shows that it’s impossible to keep the public happy. HN likes to rail on all these big tech companies amassing and abusing data, and here as they go deleting one of the most Orwellian parts it gets complaints.

It's not that there's one less tool. It's that the tool given to begin with was wrong for that purpose, and instead of leveraging the technology in beneficial ways, they're removing it.

They could have implemented in the account settings somewhere where you can let it alert you for any such usage of your face, but without it doing any auto-tagging or other publicly visible things which we usually complain about.

Knowing outlines of how online harassment happens, I'm confident it would do far more harm than help in this context. Weaponized autism is real.

HN is bizarre sometimes. When this feature first appeared, everyone acted like it was the coming of the Antichrist. Now it’s going away and people have “mixed feelings”. I think people just love to hate big tech no matter what they’re doing.

Or... maybe HN is composed of different people with different viewpoints.

People are more motivated to comment when they're upset about something, so it ends up looking like everyone here hates everything big tech does, when it may be that a small number of people hate decision X, a small number of people hate decision Y, and most people either like them or don't care.

You often hear complaints that big tech drives people to extreme viewpoints through algorithms. But the natural human algorithm of putting more effort into the negative than the positive seems to lead to the same outcome.

This is also why it's much more valuable to look at user metrics than actual feedback when making design decisions.

Individual pieces of feedback can be valuable, but if you let it do the driving then you can end up making changes that worsen your product. You become subject to the will of the most vocal critics at any given moment.

Be careful with this advice, though. Writing off feedback, as Mozilla does, can blow up in your face as you dismiss any and all criticism as being from a vocal minority of whiners.

> Or... maybe HN is composed of different people with different viewpoints.

A useful concept for people to remember as it applies everywhere. Particularly as we get drawn into tribal politics, it would be better for people to remember that the opposition is not uniform and consists of many different viewpoints. For someone who wants to be divisive, it is trivial to cherry pick examples of bad behavior ("hypocritical") by pretending that those you disagree with must all share exactly the same priorities and opinions.

Why pretend? If you genuinely believe it, the confirmation bias is likely to keep you from thinking about it enough to assess whether it's true. (Do you regularly re-assess whether the sky is blue, or whether you're wearing clothes? It's just there, background knowledge.)

You are probably right, many people aren't pretending. They really do think their opponents are a monoculture. Convenient, I suppose, critical thinking takes effort.

Yeah. None of my opponents do critical thinking; it's a good thing my lot do.

This. The biggest example I can think of to prove your point is Microsoft.

To provide a personal anecdote; I don't like using Microsoft products and always look for an open-source, unixy replacement when at home. I complain about Microsoft (products, economy, & policy) at almost every opportunity. And in the same breath I will wholeheartedly recommend the same Microsoft product that I personally wouldn't use to a client in need. The ubiquity and quality of Microsoft products is too significant to ignore. If I were to recommend one of my neckbeard alternatives I would certainly get support calls about it all the time.

But if you only ever went by my social media posts you would think I absolutely hate Microsoft and vehemently refuse to use their products. That's not the case. I just don't get excited enough about Microsoft as a business to go parroting their marketing and waving their flag. The good balances the bad enough that I stay with the brand, but human negativity bias still exists and I've got to complain. The product is too ubiquitous to ignore.

Or, you know, these are different people with different values and different sets of experience.

And also, I'm sure the goalposts have moved a lot in the 3-5 years since this feature was announced.

At the time, facial recognition was almost sci-fi levels of absurd to envision, especially at scale. Nowadays, anyone can do their own facial recognition on laptop-level GPUs from freely-available open source software and every site that can use it, probably does use it in some form.

It's a lot more normalized now than it was.

Actually your comment is pretty bizarre.

You're distilling your impressions from thousands of people into a single sentence, and then you're taking a single person's response and using that to push your agenda that "people just love to hate big tech". It's fascinating that you think this is a valid observation on your part.

Is there any one single representative of "HN" as a body of opinions?

I think assuming malice on the part of Zuck is a good default. It's as in that blockbuster movie "no matter who wins, you lose". The only winning move is not to play.

lmao

I don't think there's a need to be confused here. Parent liked it when the tool empowered them, but most people don't like when tools give others power - potentially power over them. The difference in like/hate is in how the tool is deployed.

If the only use of facial recognition on Facebook were to track how one's own image is used I don't think many people would be upset. But allowing someone else to find every image one appears in is a different and scarier thing.

>I think people just love to hate big tech no matter what they’re doing.

Or, maybe companies have proven again and again and again and again and again that not only can they not be trusted with your data, but they don't secure it, because profits.

Utilitarian arguments are a bad base for a discussion that is primarily about ethics.

> The facial recognition feature is how I figured out someone was doxxing and harassing me on their page.

What would you do if you think that somebody who harassed, and doxed you were working at Facebook, or Google?

So sad that we can't have nice things. I like the feature that automatically tags people in pictures.

Sometimes I have trouble remembering people's names (not face blindness or similar, just regular trouble I guess). I would love to have smart glasses that show the name of someone I've met before but don't quite remember. But alas, our society being what it is people will abuse facial recognition, so I can't have that it seems...

I didn't like that feature, and neither did my circle, back when I was on Facebook. We found it unnerving. But we are older and perhaps prefer the illusion of privacy, and not anyone being able to out us by uploading a picture and being archived/cataloged non-consensually.

The question is are you ok with some stranger adding a picture of you on facebook and then facebook promptly tags you on their profile.

Knowing that someone I don't know uploaded a photo of me seems like a good feature, yes.

But now everyone else knows, too. That's not such a great feature.

But that isn’t how the feature worked

Not if you set Facebook to require your approval for all tags. Then you know before everyone else.

Which in turn requires you to have a facebook account.

I’m pretty sure it was only applied on Facebook friends posts, ideally those are not strangers.

> I would love to have smart glasses that show the name of someone I've met before but don't quite remember.

Or you could just tell them you don't remember. It's rarely rude. It's better to be human. We all forget stuff all the time. If we stopped imagining that everyone else is holding everyone else to impossible levels of "quality" we'd all be much better off. Just admit you don't remember their name. If you feel really bad about it, work in a little story about what you do remember about them. If you don't remember anything, then the meeting wasn't important to you so just let it go. If it was somehow important to them, there'll be a way to resolve this. Or there won't and you'll have a lost a friend but all of that is better than hiding your humanity behind a robot.

Perhaps we need to issue a new "Turning Test". Can a human tell the difference between a human and a human with machine assistance? And does it matter?

> Or you could just tell them you don't remember. It's rarely rude. It's better to be human. We all forget stuff all the time.

I've decided that I must be atypical around name recall for some reason or another. I often do ask, because as you said it is human to forget, but I find myself forgetting like 3 times in the course of a conversation sometimes, I could really use a machine of this sort.

Great! But if it is not without the consent of the person involved though, it is your "convenience" breaching someone else's privacy. I think it is important to think about this.

I don’t see how organic facial recognition is any worse than electronic recognition. Evidently some people can take a simple wanted or missing persons person and solve a case. But for me that ability feels like magic. How is having an aid, like facial recognition glasses, any worse?

This of course assumes a non-scalable offline system. Obviously something online like China’s social credit system is more of a breach.

I am definitely anti facial recognition in general, precisely because of your second point. If I felt there were reasonable incentives for recognition to stay offline/individual scale that would be fine with me. But obviously we’re heading in the opposite direction, where corporations use facial recognition to gather even more data on users for ad dollars, and governments use it to surveil citizens for “national security.”

I don’t want to help either of those initiatives, so for now I’m strictly anti facial recognition, even for benign use cases at the micro level.

Is it really breaching someone's privacy to assist your own memory of the things they've chosen to (repeatedly) tell you?

I think the breach of privacy is giving an almost trillion dollar corporation free training data via my face and name combination.

I don’t mind an individual knowing this information, but I’d rather that individual not hand that data over to a megacorp that has a million other unknown uses for it.

This is true. But in your point I assume you are talking about the company who is doing the facial recognition I assume. But that is not the case really. Even if the person uses a home grown tech, the person can eventually share it to another dataset.

A good example is me refusing to share my photo with my friends for that app which turns you into an old guy. Another friend just out of spite shared a personal photo he has of me. Now my photo is in this Russian database which does data mining.

For better privacy, you being careful is not enough. You need to make sure everyone around you are privacy oriented. And we need to make systems that make that happen. Opt in privacy is not effective.

I’m not sure which part of my comment you’re replying to, I think we’re in total agreement.

Face blindness probably operates on a spectrum. I too aren’t face blind, but often struggle to remember the names of infrequent contacts until I get some other prompt, like vocal recognition.

I assume this is less of a genetic problem, but related to having undiagnosed myopia as a child when perhaps facial recognition pathways are in a critical development stage.

I think for me it is genetic, my dad is also very bad at this - although he is more confidently wrong than I am.

For sure it could be genetic for a subset of people — I unintentionally ruled that out.

> So sad that we can't have nice things.

Privacy is also a nice thing.

I've also been eagerly awaiting a feature that'd notify me if I ever showed up in a stranger's photos (at concerts, travelling, etc). I know FB has the technology but the feature apparently never materialized.

There's a lot of useful stuff you can do when people stop abusing that stuff.

pimeyes can do this

> I like the feature that automatically tags people in pictures.

Yeah, when it first launched I thought "this is what facial recognition is for," completely oblivious to the privacy concerns.

In a world where privacy wasn't exploited:

A photo album that automatically finds you in other people's photos is really awesome. Imagine you went to a concert and your phone died, there would likely be dozens of photos with you in the frame. You could likely amass a treasure trove of unexpected perspectives of your memories, from complete strangers.

Fun ideas that are completely incompatible with reality...

It would be fine if that information stayed with you.

The fact that it is pumped around from service to service is what causes the problems.

It seems this is likely a result of the 2019 FTC Consent Order on Facebook [1]. Facebook has had to implement privacy processes which are monitored by a 3rd party who reports to a judge and facial recognition was one of the issues sparking the order.

The consent agreement [2] has whole sections and rules specifically for facial recognition data.

Likely a review program and criteria (approved by the monitor) were set up based on these rules, and when the tagging feature was pushed through this review it failed. Rather than risk the judge sanctioning them for failing to meet the order, they decided shut down the feature.

[1] https://www.ftc.gov/news-events/press-releases/2019/07/ftc-i...

[2] https://www.ftc.gov/system/files/documents/cases/c4365facebo...

Thank you. Yes... this is almost certainly happening because of the intersection of the Illinois settlement, the current microscope FB (no, I will not refer to them as M*a) finds themselves under, and the Consent Order that they have almost certainly violated dozens if not hundreds of times in the last few years. They've done the risk assessment, and they're reducing the risk profile in order to decrease potential liabilities and penalties.

> (no, I will not refer to them as M*a)

Surely, if anything, that should be "M&a".

Serious question: I kind of liked how Facebook found people in my photos. Is there room for such a feature in a privacy conscious world?

Subjective answer: Any data that has the potential for abuse regarding privacy will eventually find its way to being abused. Providing said feature requires curating the dataset necessary to power it. How can we trust companies beholden to quarterly earnings to do the right thing when financial motivations inform 100% of their actions? We can't.

Unless regulation is brought, they're going to do it anyway. The companies are absolutely going to be doing facial recognition and use that data, they just will be doing it without exposing the use of it to us.

> Any data that has the potential for abuse regarding privacy will eventually find its way to being abused.

Right, and it is our duty to not just roll over and let it be abused. We live in the world we create. Just shrugging will guarantee it will be abused.

You need to ask the other question: Do they like it when HB tags them in yours and everyone else's photos?

I mean, they opted in (it's off by default) so I would think so.

iOS does this on-device, you can access the SQLite database containing the metadata using libimobiledevice's ifuse(1).

https://libimobiledevice.org/

The privacy cat is already out of the bag once someone uploads a photo of you without your consent. That's the root behavior that should be corrected, and it's hard because the onus is on the privacy-conscious user, to get everyone else's cooperation. The facial recognition just adds even more privacy-questionable metadata.

My dream facial recognition feature for Facebook would be to allow me, as a non-user of Facebook, to upload some hash of my face and direct Facebook to disallow any user from uploading a picture with my face in it.

That's kind of a radical position. People shouldn't take photos of crowds in public places? You want veto power over that?

In the USA at least, there is no legal expectation of privacy on a public street. So, for news reporting purposes, you can publish photos of a crowd in public. However, as soon as you step onto private property, this changes. Even at places like shopping malls and stores, the owner of that place can take your photo for security, but cannot publish it, and it is not kosher for photographers to take photos of customers in these settings and publish them without getting a signed model release form the subject of the photo. Likewise, a photographer should get explicit permission for any other private spaces such as homes.

However, as soon as you step onto private property, this changes. Even at places like shopping malls and stores, the owner of that place can take your photo for security, but cannot publish it, and it is not kosher for photographers to take photos of customers in these settings and publish them without getting a signed model release form the subject of the photo.

You're on the right track, but it's somewhat more complicated than that. There are plenty of times when you can be on private property and it's perfectly legal for someone to publish your photo.

Two companies I've worked for have brought in lawyers for staff training on this. It's nuanced.

> In the USA at least, there is no legal expectation of privacy on a public street.

Depends on the state. In Massachusetts, for example, expectation of privacy doesn't matter. What matters is whether the recording was secret or not. Note that "secret" doesn't mean "the subject knew", it means "the subject could have known." If you're walking around with a zoom lens and a camera taking pictures of someone a block away who has no idea? Not secret. If you're in front of me with a camera hidden in your tie pin? Secret.

Security camera on the side of the building? Not secret. Security camera hidden on the side of the building made to look like a bird's nest? Secret.

> So, for news reporting purposes, you can publish photos of a crowd in public.

There's generally much wider latitude for use of a person's image in relation to press coverage of a newsworthy event.

> However, as soon as you step onto private property, this changes.

Not in Massachusetts, where secretly recording someone is illegal, period.

> it is not kosher for photographers to take photos of customers in these settings and publish them without getting a signed model release form the subject of the photo

Publishing is not a very relevant term. People have a right to control "commercial use" of their image.

Not radical enough. It should be an opt-in rather opt-out. A public website should verify that every person on the photo consented to being publicly displayed. Otherwise their faces should be pixelated.

I can certainly see things going that way in EU within a decade.

I don't want every public website (and who else?) to have access to the perceptual hash of my face to identify it in uploaded pictures to ask me for opt-in.

You could just cut out the face from the picture.

the privacy has been broken once you've been photographed without your consent.

The same features for name association can be done offline once the picture exists. Just show people the picture, and see if anyone knows who it is

Consent is not required in a public setting.

Not exactly. From the perspective of building data of concern, simply uploading photos won't do that. Uploading photos just potentially creates databases of rando faces.

It is the tying of these faces to a name (identity) that's the challenge. Further, the crucial data comes from a human's act of tagging a person in a photo.

In theory one could build a perceptual hash that had a universal opt-in. But I’m skeptical such a scheme could avoid collisions (e.g. identical twins) or perhaps even doppelgänger-like strangers.

Is facial recognition so good that it can distinguish you from a billion other people? I assumed the photo recognition relied on the social graph i.e. to look for your friends or friends of friends in your pictures.

Well, honestly, you shouldn't be giving other peoples' likenesses to big corporations under worldwide non-exclusive license agreements without a signed model release consent form. However, enforcing this is almost impossible.

But really, morally, you should never give anyone else's face to a corporation without their explicitly requesting you to upload the photo to a site.

Yes, there is. Instead of publicly auto-tagging you on the photo, FB could notify you that someone had uploaded a photo of you, and offer you to either:

1. Tag yourself on the photo

2. Censor yourself from the photo (blur, pixelate, black rectangle, whatever)

Voilá, now the same feature is used to give control over your data.

Here is an interesting question. Is it at that point your data? Wouldn't it be the data of who ever took the photo? Why would it belong to you just because you are in the photo?

I think giving users power over auto tagging make sense. But giving users power over other people's photos is a terrible idea.

I think it’s basic courtesy to ask someone if they want their picture taken or not. This would only automate the process.

That wouldn't be possible in many cases. True courtesy would be paying them.

Yes, if you do it client side. I could see this working as a browser plug in.

Their announcement is at https://about.fb.com/news/2021/11/update-on-use-of-face-reco....

(via https://news.ycombinator.com/item?id=29084199, but we merged that thread hither)

> We need to weigh the positive use cases for facial recognition against growing societal concerns, especially as regulators have yet to provide clear rules.

Disclaimer: I have never owned an Apple product, and I haven't used Facebook for roughly a decade.

Apple have been doing face recognition for a while now, arguably at a similar scale, but their use case is pretty straightforward. With FB face recognition is not the real concern, it's the fact they have been caught misusing personal data multiple times.

Sadly apple are not angels either when it comes to scanning photos:

https://www.pcmag.com/news/apple-to-delay-implementing-anti-...

Where does apple do it? They do it in iPhoto but it is based on what you tell it, and is opt out. Where else?

I only know about FaceID.

Unlike Facebook's facial recognition, other people can't enroll you into Apple's Face ID without your consent or knowledge.

Face ID data is stored in an encrypted section of the iPhone's hardware and not transmitted:

https://support.apple.com/en-us/HT208108 https://support.apple.com/en-au/guide/security/sec59b0b31ff/...

Here's where Facebook's brand reputation sits in my mind today:

I do not believe them at all. I do not believe they've had some change of heart regarding "many concerns about the place of facial recognition technology in society." I do not believe they intend to not use this technology in some other privacy-invading way.

Trying to figure out their real angle here. Avoid regulation? "Good" PR at a time when they desperately need it? A broader existential threat that I can't even fathom at the level of power their company currently wields in society?

Probably not much at this point they could do to change my absolute negative opinion of the company though, so there's that.

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.

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

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.

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.

> 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.

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.

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.

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

I get all the privacy concerns, but this was actually super useful for people who opted-in. Manually retyping everybody's name when tagging them in 50 photos from an event is torturous.

Yes, but why tag people?

So they can see themselves in photos? So you don't have to look through all of your friends' photos to find which ones you are in and repost them yourself.

Isn't it possible to tag people even if they don't have a FB account?

You can type whatever you want as a tag, although I've never actually seen someone do this.

I guess it could just be used to annotate images, or to recall who's who for your own use, a la how people used to write everyone's names on the back of photos.

Who opted in? I certainly don't remember ever doing it, yet auto tagging has always worked for me (both when I am tagging pictures and when others are tagging me).

> Who opted in?

It was one of those "by continuing to use our service(s), you are opting-in" t&C's updates.

So in short: nobody did but legally there is a box ticked.

The opt-in is at https://www.facebook.com/settings?tab=facerec -- I definitely remember intentionally opting in when it was added, and it's off on some of my throwaway accounts so I assume it wasn't auto-checked for everyone at some point.

Surely one should be able to just tag _the album_ instead of each photo?

I always figured this constraint was kept as a happy convenience so that Facebook have a steady supply of training data for their face recognition service.

If they are shutting that service down, maybe they'll let us tag albums?

My wedding album has like 200 photos in it. I definitely had relatives who didn't want to look through them all and just asked me to send them the photos with them in it.

I doubt Facebook won't offer some unsupervised face clustering for albums, they will just remove tagging from images. Google Photos offers me pictures of distinctive people found in my gallery without giving me their names.

the ideas aren't mutually exclusive.

And then each guest has to go through every picture in the album to find the ones with them to share?

The hoops they have to jump through just to enable some really cool features down the line is sad.

I don't know why adding another opt in wasn't enough, it's not like the haters are going to believe they deleted the templates anyways.

“Haters” ok. I feel like that is just a cheap deflection from legitimate criticism.

Why would they? I bet most users actually like that facebook finds their face in friends photos.

I bet Facebook has a much better idea of what most of their users like or dislike.

Its probably because Facebook finally realized when government etc ask for data they need to give them otherwise you know ....... Also this is also good opportunity to green wash. It doesn't make a single dent on their ad revenue so why even gather more hate by doing some futile task?

Pity. I liked the feature. Fortunately, we’ve mostly all moved over to use Google’s shared albums with automatic facial recognition and sharing. My friends all have photos I take of them and likewise they share theirs with me.

I think Google has figured out a lot of fun sharing tools for people in trusted environments: location sharing, photo tagging, etc.

I’m rather impressed with it all. Good stuff.

Same, though Google doesn't enable it for users in different countries. It's weird to be able to automatically share photos with a partner with face recognition, but then they can't reciprocate for some reason.

A legitimate question I’ve always had when we read such announcements is how can we be certain they’re actually doing whatever they’re claiming / planning to do. Unlike software /OS modifications, it’s not feasible to somehow reverse engineer FB’s backend and find out if anything unusual is happening. At least without any attempt of compromising it.

Considering how powerful these orgs are, with low probability of facing any legal repercussions about potentially false claims, do we have any other option than to take their word for it?

The same way that you verify that the local chicken plant is properly following contamination prevention procedures. You don’t, at least not personally. We as a society rely on government based regulations and knowledgeable inspectors to verify these things. I’d say software should be no different in these regards. (Though that isn’t currently the case.)

Reminds me of the drone photographers that were flying over pig farms in the south that were egregiously breaking laws. Instead of fixing the pig farms, they went after the photographers.

Authoritarian dysfunction. Killing the messenger.

Nowadays there's enough dissent inside FB itself (according to news articles), that if they lied about this, someone from the inside will probably leak/whistleblow about it...

They are finally realizing that they're scaring people away.

..after the Horse has bolted!

2021 is the “let’s bring back normal stuff” year, after Apple restored USB, HDMI, Magsafe ports and the F1 keys ;)

What next, moving back from the Cloud, data ownership, Youtube having chronological ordering instead of “We chose what you should watch because we want to influence your political opinions” ???

>Youtube having chronological ordering

I am confused, for me the subscription section is always chronological and has everything I subscribed , I see the youtube home page a few times a year because I bookmarked the subscription url and not the home page.

https://www.youtube.com/feed/subscriptions

Make Blogs Great Again

Meta has no plans to discontinue the use of facial recognition though: https://www.vox.com/recode/22761598/facebook-facial-recognit...

The story of how Facebook's facial recognition capabilities came to be is an interesting one: https://en.wikipedia.org/wiki/DeepFace https://en.wikipedia.org/wiki/Face.com

I had no idea that Face.com's algorithm was/is able to ID ~97% of all faces 'in the wild'. Nor did I know that Face.com had opened its API to use by dozens (hundreds?) of companies back in 2007. I wouldn't doubt if there are knock-offs / clones / reverse-engineered versions of the original Face.com API floating around. API security and design - in 200x - wasn't what it is today.

They're not going to shut it down. They're just doing to deny access to you and me, and sell it to cops and despots.

Highly unlikely that they will delete it. The info collected is too valuable for them to delete it. They may retain the hash or other unique identifiers even if they delete originals. They have shown again and again that they can't be trusted.

trust, once lost, can never be regained.

Getting rid of its facial recognition system is a good step, but lest people either forget or miss...

  Although Facebook plans to delete more than one billion facial recognition templates, which are digital scans of facial features, by December,

it will not eliminate the software that powers the system,

  which is an advanced algorithm called DeepFace.

The company has also not ruled out incorporating facial recognition technology into future products,

  Mr. Grosse said.

Also: https://www.washingtonpost.com/technology/2021/11/02/faceboo...

https://archive.md/JMwS2

Sorry for me not believing this. They’ve been caught lying so many times trying to save face that I fully expect them to double down on facial recognition after this announcement.

If it had been opt-in from the getgo, I think it would have been less controversial.

I still remember a time when it was common courtesy to ask someone before you took a picture of them.

Facebook facial recognition never actually learnt how to differentiate my face from my relative's face. It was cool since I had no "tracking" for them to save.

Its public facial recognition system? For some reason I doubt that they will stop using it for themselves.

The govt uses facial recognition for unemployment benefits because of the fraud experienced.

I am for this decision, but I don't understand their justification for it. What was the actual reason they pulled back on this? The US government is assuredly for maximum surveillance, so long as the publicly visible part is done by private companies that can be subpoenaed for their data. FB needs to stay in the good graces of the military-industrial-political complex to avoid anti-trust action.

FB doesn't do anything out of the good of their heart, so what are they getting out of this? I really doubt that winning the hearts of civil libertarians is at all important to them.

(I also noticed how the NYT managed to get a dig into the discredited China story when discussing a company that was surveilling billions of people worldwide using facial recognition technology.)

This is all just PR. Doesn’t really affect revenue.

a step in the right direction. that said, i closed my facebook in 2008 and i never missed it.

Good riddance. Good choice by FB.

I’m happy but I don’t believe it.

How about Match Group?

It's always confusing to see what the internet police go after.

Total scams ripping thousands to hundreds of thousands off from people (tech support scams etc). Crickets.

An opt in feature on the facebook platform that is useful (at least to some). Multiple Attorney General press conferences :)

There's Jim Browning for that

But their facial recognition is not opt-in. You cannot prevent other people from uploading photos of you then tagging your face in them, training Facebook's facial recognition to your face even if you consciously never upload your image a single time.

If you don't like that, you're going to hate this.

https://youtu.be/KT18KJouHWg

Basically they can find you from your DNA, even if only a couple very distant relatives ever gave theirs.

I get what you're saying but that's unguided learning though.

Facebook doesn't have any way of linking that face fingerprint to you to even check your consent - but a part of me doubts that when they do make the link they remove the fingerprint bc you didn't opt in.

You can only be suggested as a name for someone else's photo if you've turned face recognition on.

"If you’d like other people to see tag suggestions for you, you’ll need to turn on your face recognition setting."

So you need to opt in to it to let others tag you.

But how does it know that "This is slownews45's face, so don't show any auto-tag recommendation for it."..?

You got the algorithm backwards, if slownews45 uploads a pic, the system gathers all of their friends who opted in and compares any found faces to the templates of the opted in folks. If it doesn't get a hit, it simply moves on

The negative impact of facial recognition technology is much bigger than the impact of scams.

Could you be more specific ?

Are we not going after both?

Let's marshall the energy. I'm sure we can find it. Wanna lead the charge?

No - we are not going after both - the AG's take high profile low hanging easy target cases.

I've personally tried to get police / DA / AG type response on a fair number of white collar type cases. Crickets.

Stolen checks / AP fraud, even stuff that would be pretty damn easy to track down (ie, whose name was on account that cashed this check) - there is almost no follow-up. It really can be eye opening.

It makes more sense when you stop and think that there are no single group of "internet police" and that in reality it's millions of people with their ethics, personal views and natural inability to rank risk.

Slightly off-topic but I deleted my Instagram yesterday. I was really holding out for one person who almost refuses to talk to me on any non-FB property but...oh well. I'm glad to be rid of it, but I can't shake that it's somewhat meaningless and I'll miss the many content creators I followed who have earned my business. I wonder if there's some way to follow them in a way not tied to my identity.

There are several Facebook and Twitter proxies; perhaps there's an Instagram proxy with an RSS feed?

What do you mean by FB/Twitter proxies? Is it what it sounds like, service to access FB/Twitter in some kind of anonymous mode? And could you name a few?

Yes. There's https://nitter.net/ for Twitter. Can't remember the Facebook one, or the other Twitter one. There's also https://teddit.net/ (and another one) for Reddit.

There are: just Google “instagram bridge”. I use one of them - a bit slow but fine otherwise.

Create a new Instagram account and follow them? Instagram is anonymous.

"We are never ever ever getting back together." Facebook user base quoting 21 century poet laureate.

Hey you remember all those times we analyzed you photos, and those times we did things we said we wouldn't? We totally aren't going to do those now. We are "Meta" not the evil Facebook, trust us. "Meta" has never lied to you, right?

Who needs facial recognition when we're going to be uploading our minds into the metaverse?

"I don't believe you" https://c.tenor.com/9PqKT8E9RNkAAAAd/gif.gif

Luddites. You can't un-invent technology. Like all technology, facial recognition has tremendous potential for good. We should learn to live with the technology, not attempt to forbid its use.

Also, Meta is a bunch of cowards nowadays. They should not have kowtowed to angry Twitter activists. It sets a bad precedent.

I don't like this system where activists bully big companies into compliance with their idiosyncratic moral systems instead of using the government and the laws to codify restrictions and rules in a transparent way that everyone can understand and in which everyone has a say.

If facial recognition is so bad, it should be illegal. If it's not illegal, FB has a responsibility to its shareholders to maximize profit within the law. If, as an activist, can't get support for making facial recognition illegal, then maybe it's not quite as bad as you claim.

"activists bully big companies"

What's up with this plainly ridiculous usage of the term "bully"? It's almost self-parody. It's something I've noticed in British newspapers sometimes where MPs will whine about being "bullied" by citizens who post rude comments online.

It's really obviously not the right term for whatever point you want to express! Taking points off your English paper for incorrect word usage.

> Also, cowards. I don't like this system where activists

Are you sure it isn't just that not enough people opted in for it to be worth the cost of maintaining that part of the service?

> FB has a responsibility to its shareholders to maximize profit within the law

Which it may be doing here, by removing infrastructure and people from running and maintaining a feature, resources that can then be used on features that have a better effect on the bottom line.

Or are you believing that it is entirely about “finding balance” morally (or at least in the court of public opinion), rather than balancing the bottom line financially (even if just in part)?

Reply on news.ycombinator.com