• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Meta Removes A.I. Feature on Instagram After Days of Backlash

philb2

2[H]4U
Joined
May 26, 2021
Messages
3,471
https://www.nytimes.com/2026/07/10/technology/meta-muse-images-instagram-removal.html

The feature, which Meta unveiled on Tuesday, automatically opted in any Instagram user with a public account. As a result, countless people’s likenesses were used in A.I. images without their consent. Users complained about privacy and copyright concerns

The release of the feature was the latest in a string of missteps by A.I. companies that allowed the technology free rein to manipulate a wide range of images.
 
Of course it should be opt in, but it's always hilarious when people complain about privacy over photos they post on the internet. It's like complaining about privacy over a photo you stapled to an utility pole down the road.
 
Of course it should be opt in, but it's always hilarious when people complain about privacy over photos they post on the internet. It's like complaining about privacy over a photo you stapled to an utility pole down the road.
At a glance it's easy to judge but consider the example of that lawyer who was facially recognized by Madison Square Garden's surveillance cameras and blocked from entry merely because her photo was on the website of the lawfirm she works for, which represented a client who years prior sued MSG.

That is multiple levels of spite for something she wasn't even directly involved with, all thanks to the the availability of data she didn't even have control over. No one could predict such an outcome.

So while some are knowingly plastering their data online I always think that it's not a choice for some in various scenarios (including those who appear in others' photos and get marked by IRL friends on such platforms), nor are the consequences easily foreseen, and so privacy protections should be a baseline to help everyone equally.
 
At a glance it's easy to judge but consider the example of that lawyer who was facially recognized by Madison Square Garden's surveillance cameras and blocked from entry merely because her photo was on the website of the lawfirm she works for, which represented a client who years prior sued MSG.

That is multiple levels of spite for something she wasn't even directly involved with, all thanks to the the availability of data she didn't even have control over. No one could predict such an outcome.

So while some are knowingly plastering their data online I always think that it's not a choice for some in various scenarios (including those who appear in others' photos and get marked by IRL friends on such platforms), nor are the consequences easily foreseen, and so privacy protections should be a baseline to help everyone equally.
That's hardly a privacy issue. That's unfair discrimination.

If someone else posts your picture online there might be a privacy issue, but only if you have a reasonable expectation of privacy. I.e. If you appear at a public event you accept that photos might be taken of you and posted online.

It's not the social media site's responsibility to protect you if your friends decide to post a picture you are visible on. Should instagram require a signed copy of a model release form for every person who is visible on a picture? That's unreasonable.

How is it exactly that you think social media should protect privacy?
 
That's hardly a privacy issue. That's unfair discrimination.

If someone else posts your picture online there might be a privacy issue, but only if you have a reasonable expectation of privacy. I.e. If you appear at a public event you accept that photos might be taken of you and posted online.

It's not the social media site's responsibility to protect you if your friends decide to post a picture you are visible on. Should instagram require a signed copy of a model release form for every person who is visible on a picture? That's unreasonable.

How is it exactly that you think social media should protect privacy?

I doubt it was the privacy issue people were most concerned about it was easily allowing people to suck your picture into an AI tool and manipulate it to their own whims. The big push came from celebrities and commercial accounts from the couple articles I read. Some companies were calling Meta directly and demanding their change their tune.
 
That's hardly a privacy issue. That's unfair discrimination.

If someone else posts your picture online there might be a privacy issue, but only if you have a reasonable expectation of privacy. I.e. If you appear at a public event you accept that photos might be taken of you and posted online.
Privacy encompasses more than just collection, it's also about opaque data processing (see Daniel Solov's "I've got nothing to hide" paper). The lawyer couldn't have foreseen a company photo she has no control over will be used to be trained to detect her likeness and then target her at a venue.

By protections I mean from regulation about consent, data handling, penalties, rather than coming from social media companies who don't have user-first goals. GDPR is an example of that direction (even with some of its flaws) and it did manage to get companies to begrudgingly implement better opt out options and data handling for users. So it shows that intervention is possible and things may not have gotten so bad with better protections.

In Germany for instance you need a person's permission to share someone's photo even if they were in a public space, if they're identifiable. Can also see this in various dashcam videos from Germany where they blur people and their plates. So default consent-requiring scenarios are possible.
 
Privacy encompasses more than just collection, it's also about opaque data processing (see Daniel Solov's "I've got nothing to hide" paper). The lawyer couldn't have foreseen a company photo she has no control over will be used to be trained to detect her likeness and then target her at a venue.
I still wouldn't call that a privacy issue, even if at some places legally it's handled under privacy. It's a data processing issue whether photos publicly available on the internet should be used in training AI, but it's already too late since they have been used like that already.
By protections I mean from regulation about consent, data handling, penalties, rather than coming from social media companies who don't have user-first goals. GDPR is an example of that direction (even with some of its flaws) and it did manage to get companies to begrudgingly implement better opt out options and data handling for users. So it shows that intervention is possible and things may not have gotten so bad with better protections.
GDPR didn't do anything, well at least anything good. It created faux opt out options that are just good to annoy you, because you can't really opt out of anything, you just have to accept their terms to be able to use the site / service. In a way it's worse because when they do misuse your data they can just wave the consent popup that you had to accept.

It also caused a bunch of smaller sites to shut down because they didn't want to deal with the legal implications of being GDPR compliant, even though I never heard of any cases where it was actually enforced.
In Germany for instance you need a person's permission to share someone's photo even if they were in a public space, if they're identifiable. Can also see this in various dashcam videos from Germany where they blur people and their plates. So default consent-requiring scenarios are possible.
The obsession over privacy in Germany comes from a political campaign in the early 2010s, it is not an organic thing and does more harm than good. They didn't even allow street view for like 10 years, but when they realized how useful it is they begged google to do it, after essentially banning them for years. The only worse place in this is Austria where you are not even allowed to have a dashcam. If you are caught with one it was like a €200 fine last I checked. Who is this really protecting besides bad drivers?
 
GDPR didn't do anything, well at least anything good.
Ignoring the positives is an oddly lopsided take. The ratio to things trying to enforce better user-first protections is slim while the default state of things is wild west of anything goes (the very things people complain about until lawsuits occur).

Take Steam refunds, it was literally only because Australia's ACCC sued them because of consumer protection laws that Valve even offers refunds for everyone (which they didn't even have to do, they just expanded the scope). They didn't support it before. Then, due to the shift in buyer expectations from Steam GOG improved their refund policy (which used to have a lot of likely deliberate friction). All due to enforcement and you don't hear anyone complaining because there wasn't big tech lobbying trying to "organically" persuade the public on that one.

We've seen similar scope-broadening with GDPR where after EU countries have enforced it services have offered the same options to users worldwide (which wasn't even a requirement). Because of GDPR's leverage there are now right to erasure options for a wide range of sites, tracking consent on sites (I could get into this separately but functional cookies don't require consent and I do think consent over options should be a browser-set option instead of arbitrary on-site banners since UX is the main complaint) which allows me to choose to decline data sharing and personalized ads with over 900 third-parties on Thingiverse.com for example, fining Facebook for bypassing GPDR for personalized ads consent and enforcing proper opt-in, etc.

I could similarly argue (and have) GPL licenses have flaws but still have positives for pushing for and variously enforcing by lawsuits user-first rights but they serve a purpose that has mitigated more corporate abuse.
 
Ignoring the positives is an oddly lopsided take. The ratio to things trying to enforce better user-first protections is slim while the default state of things is wild west of anything goes (the very things people complain about until lawsuits occur).
I'm not ignoring anything, I simply never encountered any positives stemming from GDPR, while I'm dealing with the negatives constantly.
Take Steam refunds, it was literally only because Australia's ACCC sued them because of consumer protection laws that Valve even offers refunds for everyone (which they didn't even have to do, they just expanded the scope). They didn't support it before. Then, due to the shift in buyer expectations from Steam GOG improved their refund policy (which used to have a lot of likely deliberate friction). All due to enforcement and you don't hear anyone complaining because there wasn't big tech lobbying trying to "organically" persuade the public on that one.
Nice backhanded accusation of me being somehow influenced by big tech. I specifically said I hate GDPR because a bunch of sites I was a regular on were shut down because of it, including a similar tech forum to H with a great community that had 20 years of history, where I had 30k posts, going back to my teens. How is that big tech influencing me? No, fuck GDPR, and I mean it.
We've seen similar scope-broadening with GDPR where after EU countries have enforced it services have offered the same options to users worldwide (which wasn't even a requirement). Because of GDPR's leverage there are now right to erasure options for a wide range of sites, tracking consent on sites (I could get into this separately but functional cookies don't require consent and I do think consent over options should be a browser-set option instead of arbitrary on-site banners since UX is the main complaint) which allows me to choose to decline data sharing and personalized ads with over 900 third-parties on Thingiverse.com for example, fining Facebook for bypassing GPDR for personalized ads consent and enforcing proper opt-in, etc.
Many sites only do malicious compliance where in order to opt out you have to click through pages upon pages of individual cookie options, and then do it again on each visit. They know people will rather just click accept. GDPR might have been well intended but it does nothing until they go after sites that engage in this type of malicious compliance and offer no easy opt out. Or worse, put the opt out option behind a paywall.

Right to erasure doesn't work in practice either, sure you have the option to close your account, but that doesn't mean they delete your data, they can cite a number of reasons why they can keep it basically indefinitely, so the only thing they do is cut off your access when you close an account, while they'll still happily sell your data to brokers.
 
Many sites only do malicious compliance where in order to opt out you have to click through pages upon pages of individual cookie options, and then do it again on each visit. They know people will rather just click accept. GDPR might have been well intended but it does nothing until they go after sites that engage in this type of malicious compliance and offer no easy opt out. Or worse, put the opt out option behind a paywall.

The Facebook fine I mentioned is an example of just that: an EU member going after Facebook for trying to skirt the guidelines and got fined almost €400m. They do go after companies using dark patterns to obscure consent, one of the very first fines was another instance of Facebook trying to do that with their permissions screens.

Brave browser found that Google was using a GDPR workaround and it got reported and was investigated for infringement. It's either something holds companies to some degree of accountability (legislation or costly lawsuits) or it's left entirely unmitigated.

Right to erasure doesn't work in practice either, sure you have the option to close your account, but that doesn't mean they delete your data, they can cite a number of reasons why they can keep it basically indefinitely, so the only thing they do is cut off your access when you close an account, while they'll still happily sell your data to brokers.

Part of GDPR is having the leverage to call them on this. France fined Discord for failure with GPDR to actually delete user data, that's yet another example of where compliance was enforced. If the response is just 'well they'll just figure out a way to do it anyway' that's not entirely incorrect for various companies but one could take a simple position like that about anything, when even lawsuits are just a best-effort to get compliance and push behavior in a different direction.

I specifically said I hate GDPR because a bunch of sites I was a regular on were shut down because of it, including a similar tech forum to H with a great community that had 20 years of history, where I had 30k posts, going back to my teens. How is that big tech influencing me? No, fuck GDPR, and I mean it.

My reply didn't say there aren't things worthy of critique, I can list flaws about any number of things I have opinions about. The UK and US' age assurance legislation has affected various small websites in regard to actual compliance and over potential liability and that's the type of legislation I strongly oppose (in particular because there are privacy-preserving ways to achieve it and they're not brought to the table).

Want to know criticism I have about GDPR's guidelines on erasure? It has caused unnecessary deletion of user content due to the clause about automatically having to delete content from inactive users (rather than on-demand), which is literally the reason why I have to log into Google and other services now bi-annually to prevent this (and in the case of Ubisoft they applied an interpretation of this even for users who have bought games, wiping their purchases).

Do I agree about that automatic aspect? Nope. I can still see other benefits though.

If we could roll back the internet to before tech companies and data brokers were trying to vacuum user data then there wouldn't have been any impetus for such regulation and the net would have remained mostly the domain of nerds. Instead we have an internet mostly influenced by shitty companies with anti-user practices and have to consider what can be done to affect change.
 
Back
Top