mysterium

Thank you for writing this.

Of all the topics I’ve had to work with in my career, this one has caused me by far the most frustration. I like to think the hacker community is generally scientifically-minded and open to rational debate, but online discussion of this subject uniquely tends to cause people to hunker down, refuse to engage productively, and resort to name-calling. This might feel righteous, but ultimately leads to own-goals from us.

Firstly, to make one thing clear, it’s _absolutely_ possible to do age verification in a privacy-preserving manner. A technology called Privacy Pass exists that separates the roles in the age-verification question. This would make it possible to have a solution where the government can attest to your age without knowing what website you’re trying to visit (e.g. pornography, or an online casino, or just purchasing alcohol online). This is just a matter of fact. I’d recommend reading RFC 9576 for more details on the separation of roles here, it’s a really nice protocol.

There seems to be some misconception that privacy-preserving solutions for age verification aren’t permitted under various legislations. I don’t know where this comes from, but certainly Ofcom _mandates_ the minimisation of unnecessary data collection. This doesn’t mean that suboptimal technologies aren’t in use, but there’s certainly nothing precluding the use of fully privacy-preserving solutions.

We should be pushing for privacy-preserving age verification. It’s easy and convenient to say it’s the job of parents, and label anyone who doesn’t use parental controls as a bad parent, but at the end of the day, a government’s job is to look after its citizens regardless of whether they have good parents. If instead of engaging productively we stonewall the topic based on a vaguely-directed-but-intense distrust of Government, then governments will implement it anyway, and the solutions will be bad. We know this is the case, because it already happens.

I participated in a very productive workshop last year with representatives from government as well as various privacy-conscious companies, including Mozilla. I was pleasantly surprised at how productive we could be when we all worked together on this. We all walked away with a much better understanding of some of the problems, some of the nuances involved, and some possible paths forward.

Characterising people who analyse reprehensible material in order to try to save children from abuse as “pedos working for government to watch all the reported images” is, quite frankly, disgusting.

The thing is, it _is_ controversial. At least amongst the general public.

Obviously not in somewhere like Hacker News where there’s a clear consensus, but if you asked a random sample of the UK population “should law enforcement be allowed to compel tech companies to hand over all DMs of confirmed paedophiles?”, I’d bet very good money the majority would say “yes”.

The notion that “Big Tech” can absolve themselves of the responsibility to help law enforcement find child abusers by saying “it’s all encrypted, not my problem”, does not sit well with a large sector of the population.

Whether it’s good or bad is an ultimately political question, and both sides of the debate tend to talk past each other on this topic, but it’s undeniably a controversial point within the broader population.

This feels like a perfectionist fallacy. One has to start somewhere, pornography is something we can pretty much all agree children shouldn’t be watching. It’s theoretically a mostly-achievable goal.

Solving one problem doesn’t mean we can’t also work on solving the others.

Very reasonable, and indeed the July deadline will make it very likely that they _don't_ use a privacy-preserving method like Privacy Pass (and it's very reasonable for people to raise concerns).

My main issue is whenever the topic of age verification is raised, the consensus from the tech community is "the only way to achieve age verification is mass surveillance and I won't stand for it!", and as someone who works in an adjacent field to this, I feel the need to dispel that notion. We should be directing our collective energy into ensuring age verification is secure and private, not fighting against the notion entirely.

Hot take: I'm in favour of this.

There's actually been significant progress in developing privacy-preserving attestation technologies in the last few years, that could enable anonymous age verification for adult websites.

There's an IETF protocol called Privacy Pass (see RFCs 9576-9578) that can attest facts about clients (e.g. their age) in an anonymising manner. You reveal your age by showing your ID to an age verification service (e.g. a government run service, or a bank), but crucially not the reason _why_ you're proving your age (i.e. the bank doesn't know if you're visiting a porn site or a casino). That service gives you a bunch of signed tokens proving you're over the age of 18/21/whatever, and you can redeem a token on the porn/gambling/alcohol-purchasing website for access. It uses various fun bits of cryptography (e.g. blind signatures) to ensure anonymity for the user.

With such technologies, anonymous age verification becomes achievable. Now, whether age verification is inherently "bad" is up for debate, but that's a matter of opinion. Obviously this technology isn't infallible and it will be possible to work around, but we don't need a perfect solution here. Preventing 99% of children that would have viewed violent pornography from viewing it is a massive win, and the fact that we can do it in a provably-privacy-preserving manner using cryptography is very very cool.