It has nothing to do with youth protection. It's pure authoritarian shit.
post
Sure you can only determine if "user is of age or not" with zero knowledge proof. But collect enough proof and statistics can de anonymize that anyway. How is no one addressing that?
Is this the one you can bypass by changing a config file?
This is being pitched as a thing for kids, but kids aren't champing at the bit to prove they're children. So, presumably this just means that any time you access a service, you're gonna have to use this to prove your age, as an adult, right?
Well, kinda. You can have access tiers. For instance no access for age<13, limited access for 13<age<18, full access for age>18.
Needless to say, I think this is (in most circumstances) the wrong approach.
We are talking about products that are often deliberately harmful and hostile to all users, and then we expect to have a child-safe version of these. Shouldn't we try to get an adult safe version too? It would be way easier to protect children then.
For sure, but in all those access tiers, they would presumably need to verify MY age as well to be sure of which tier I fit into, right? The only reason I can see that I wouldn't need to use this app is if I'm on a site that isn't age gated, I think.
Edit: sorry, I didn't even respond to your second point. I think you're exactly right about the issue being with them being harmful to begin with. Fixing those issues should absolutely be the first priority.
And I wanna be clear, the idea of a zero proof system is the best I've heard of so far in all of these concepts. At least with that there aren't some insane security issues being overlooked. I'm just wary of the surveillance state future everyone seems to be pushing for.
Assuming the technical implementation is sound (I'm techy but that's still way over my head), there is something missing from the explanations I've been seeing so far.
The state is of course the one who should be proving my identity, and the website has (usually) no business knowing who I am or holding a copy of my documents. The state however has no business knowing what I'm browsing, and a pinky promise is not enough.
I can't understand whether this is something that the proposed system offers, or whether it's a property of zero-proof systems in general.
Obviously something like this must necessarily be Free and Open Source if any trust at all must be put into it.
and a pinky promise is not enough.
Yah that's my sticking point too!
I believe that under good faith, Zero Knowledge Proof could work and guard privacy from both the gov and the sites.
But "good faith" is doing heavy lifting. The desire to corrupt the system in some way that turns ZKP into secretly non-ZKP is going to be huge. Even if it begins OK, we will all become locked into it. And if it gets corrupted years later, too bad so sad, because we're locked into it!
We've already seen intelligence agencies trying to corrupt encryption standards, to look secure when they have a secret flaw. That's the kind of corruption I worry about with ZKP age gates.
You cannot turn a ZKP into being secretely not ZKP without significant effort though.
Take the following example protocol:
- Social media app sends you a token to verify.
- You append a private secrete string to the token and hash it with a known, collision resistant hash function.
- You send the hash to the government's server and request an "18+" signature. The signature should correspond to a public key.
- You send the signature back to the social media app, including the secret you used.
- The social media calculates the hash of its token + your secret and then checks whether the governmen't signature is valid with that value.
The government will not know which social media site was used, the social media site will not discover anything about your identity beyond a binary "is above 18 years old" statement. This is because you control all communication.
To discover anything else, they would BOTH have to collude in some significant way. They can only do so in step 5, by having the social media app send the value you gave it to the government. Maybe there exists a protocol that you control that works against this threat as well, I'm not sure.
But if they collude in step 5 - what prevents the social media company from sending all information it has about you to the government already all the time, even without age verification? Like IP addresses, phone number, access time etc. If the government further controls all the ISP servers and log which traffic from where goes where, it could certainly identify you already.
First, I wanna say I appreciate your reply. It's well made. I believe you, mathematically, about how ZKP's work.
I just think that when rubber meet road, there will be potholes. Example, strong encryption cannot be broken, practically speaking. The social media companies make real E2EE. But they control the client. So they simply scrape post decryption from the user's device. It's true, the E2EE was secure. But that didn't matter in the end. There was a way to circumvent.
We'll see about ways like that with ZKP. I'm not smart enough to know how it may happen. Only that the incentives will be big. Encryption isn't defeated by breaking the math. Neither ZKP. It'll be some other way. Something sleazy.
the social media site will not discover anything about your identity beyond a binary “is above 18 years old” statement.
To discover anything else, they would BOTH have to collude in some significant way.
I would say, social media can already discover most ppl's identity. Without having to collude at all. There's a whole ass industry of identity resolution, even when ppl don't mean to give their own identity. Would social medias stop doing that, just because now ZKPs? I'm afraid it may deliver a false feel of security.
Can you explain a little what you mean with:
So they simply scrape post decryption from the user's device.
As far as I know, no social media company's posts are E2EE. After all: It's not possible to have both public posts and E2EE. "Direct messages" to other users can be E2EE but you'd have to trust the company with the encryption keys.
The only condition that requires Zero-Knowledge Protocols to function is that your device is not hijacked by hackers (and there are no deliberate backdoors and such). This can be achieved by having the app be open source with regular security audits. The social media company can do nothing to identify you, nor could the government (unless again, they collude and share secrets).
But yeah, social media can already identify most users because of surveillance capitalism. The goal however is to ensure identification is not in any way made easier via age verification.
As far as I know, no social media company’s posts are E2EE
Oh, sorry, I shoulda been specific. I meant cases like the Whatsapp class action. It hasn't been proven in court yet, tho. The prosecution says they have evidence Whatsapp and Meta get access to E2EE messages since they control the client. But it's also important to say that Meta is denying they do this. It will play out in court.
The 52-page privacy lawsuit contends that although Meta has pervasively marketed WhatsApp as a private, secure, end-to-end-encrypted messaging service where “[o]nly you and the person you’re talking to can read or listen to” messages, Meta employees, Irish consulting and tech firm Accenture and possibly other third parties, unbeknownst to users, can access messages via a “backdoor” in the WhatsApp source code. According to the complaint, the backdoor allows Meta and WhatsApp employees and/or third-party contractors to “circumvent the encryption in order to view users’ private messages.”
Whether it's proven, or not, it's possible to do it. The mesages have to be decrypted for view.
and there are no deliberate backdoors and such
Agree, but that condition is doing some heavy lifting. As in the alleged Whatsapp case.
Yeah, that's again a little different. You have to fully trust WhatsApp that they are doing what they promise. You can't really verify this yourself.
Besides, if the app is open source, backdoors are generally more difficult to implement. Especially for something like E2EE, where people look very closely at what the application does with keys. Same with age verification in my opinion. You'd need to pull off a lot of gymnastics to put in a backdoor, see the xz utils one which was only achieved through several obfuscated stages in a codebase rarely ever looked at by another human.
Well, it depends what you mean about "adult safe".
There are still regulations to make sites free of harmful content. Problem is that is not enforced enough.
But do you want governments to decide what's "adult safe"? It seems that's an even worse approach with far more overreach than already proposed.
Adult safe should mean porn ban or explicit content, if you want to make unified rules for all ages.
We should also take into account that there factually is a difference between kid brains and adults. This is not coming from nothing, even though the proposed solution ends up being a privacy nightmare.
Surprise, the app is insecure hot trash: https://xcancel.com/Paul_Reviews/status/2044723123287666921
This is repression.
Minorities (especially trans and/or neurodiverse children, who often aren't even respected by their family) will be especially affected by this censorship of their only safe space - the rare inclusive corners of social media (like the fediverse and some rare mainstream social media communities).
If I hadn't used social media in my childhood, I would be convinced what my parents and the Slovenian medical system tried to implant in my mind: that all autistics are idiots, and that I shouldn't be educated, I should just have a job in some kind of disability institution.
Maybe this is the intention of this legislation... Also, is the app still deliberately incompatible with Graphene OS via strong-integrity? Hmmm...
It is not hard to avoid social media and to keep kids off it. Most parents just don't want to. And children are less academically inclined because of current culture not technology. Without this tech but culture being as it is now would result in the same thing.
Kinda fucked up honestly, companies paying billions for algorithms to maximise engagement and the government goes for the victims instead of the companies.
I am not sure if you are referencing what I am saying or adding on to it. I absolutely do think the social media mega corps should be held accountable but it is not because children are using their platform. They are abusing everyone on their platforms and that should be punished/stopped.
"Unable to provide effective mechanisms"
Why don't we replace with unwilling because they sure as shit do. Not. Care.
How would that work? How would you distinguish the "pedos" visiting from anyone else? Whatever these "children specific sites" you're referring to, are children going to create, host, and manage them all on their own without any adults ever to help them at any level?
all 4 comments