The purpose of a system is what it does
There’s a hit tweet that goes as follows:
My neighbor told me coyotes keep eating his outdoor cats so I asked how many cats he has and he said he just goes to the shelter and gets a new cat afterwards so I said it sounds like he’s just feeding shelter cats to coyotes and then his daughter started crying.
Wikipedia has an article on this concept, titled The purpose of a system is what it does. It’s a cynical and seemingly irrational attitude, but at some point you have to look at the functioning of a system and ask whether you’ve built a coyote-feeding machine.
Yes, I’m back on my anti-censorship rant again.
The failure of age verification
I’ve previously ranted about the Online Safety Act, a British law which is ostensibly about child safety, but has disastrous side-effects for everyone else.
In short, the OSA ignores the plethora of existing online safety tools, especially parental controls, which are so well trusted that parents felt safe allowing unsupvervised internet access to an entire generation of children, the “iPad babies”. Instead, OSA requires all existing websites, the vast majority of which are based outside the UK and cannot be forced to comply, to implement age verification.
Only a very limited set of verification methods are permitted, all of which violate the individual’s privacy. They are cumbersome enough to discourage users, and expensive enough for the site operator at scale to undermine the business model of free ad-supported content.
One has to wonder if that’s the point.
Proof of age
Suppose a British man wants to buy alcohol. He carries his purchase to the self-service checkout and scans it, where a staff member glances to confirm the gentleman’s appearance before swiping approval. No record is made of the buyer, nor any imposition made upon him. His counterpart in the United States is more likely to be made to show ID (such is the cost of living in the land of the free), but even then the cashier only glances at the date field, which he has forgotten by the time the customer has left the store.
Next, he wants to visit a website featuring adult content—a much safer pastime, given that alcohol kills nearly 10,000 people in Britain each year (source). Our man has already disabled the built-in adult content filter on his wi-fi hub or at his phone network, and last week was forced to present government identification papers to continue using his own iPhone. Today, he faces new inconvenience: he must create an account on the adult site to prove his age.
Now comes the humiliation ritual. He must switch on his device camera and make a little performance. Turn your head on command so we know you’re not a recording—the provider is fully aware of the potential for such a bypass. Open your mouth. Close your mouth. Open your mouth again so you can watch porn.
Who would buy alcohol in the supermarket if that were the harrowing procedure?
Failure states of the system
A gating system fails in one of two ways: when it fails to deny someone who should not have access, and when it fails to allow someone who should have access.
Fail deny modes
A major failure is when the system locks out the people it was intended to let in; in this case adults. Going foward, as legislation passes in other countries, this will affect more people, as well as non-adult services like social media and computer operating systems in general. (Remember that adult content is often used as a wedge in censorship and control, as fewer people will openly defend it; the legal precedents it sets and systems it creates will soon come for everyone else.)
The first failure mode is lack of support. Most sites implement only one of three approved methods: camera-based, government ID (or combination of camera and ID), or credit card. Most use a third party verification service, who in most instances is some new startup not backed by a trustworthy brand.
The second is that implementing camera or ID kills the operating model or financial viability of many sites. Camera/ID verification current costs thirty cents, and is reliant on the continuing cheap availability of AI or underpaid third-world workers, neither of which are guaranteed going forward. Video advertising pays in the range of 0.1 to 0.5 US cents per view. Transactional verification is non-viable for ad-supported sites, so users have to sign up.
Any obstacle to signup on a site also decreases engagement. Recall the professor who saw reduced sales after moving his book from Amazon to a site where most people didn’t already have accounts. The intentionally signup-free experience was one reason 4chan grew faster than competitors, and once appeared in Alexa’s top 100 US sites. Recall how when Amazon was an online bookstore, free shipping resulted in increased sales, except in France, where the shipping 1 Franc (US $0.10); the psychological distance between free and any amount of money is huge.
Even credit card payment imposes a barrier to entry. Credit cards are the least privacy-harmful option by a small margin, but are only readily available to the successful in our society. If you are poor, unemployed, or have no credit history, you cannot acquire a credit card, and therefore cannot use any site which uses credit card authentication only. This is bad for you and for the site.
Another huge category of failure modes is when the site can reliably determine that an individual is an adult, but by means not approved by the government. Account age is forbidden as an identifier. Zero-knowledge proofs were known of when the OSA was drafted, but not included. Each site pays the verification fee each time, rather than consulting a common service; a service could validate and sign a PGP key or similar, but this was not considered.
Fail allow modes
The real stinker is that age verification is so trivially bypassed that it doesn’t even fulfil its stated purpose.
Some sites will not comply, to begin with. We must therefore continue to use the existing parental control systems, rendering the age verification systems redundant.
Even so, all three major ID methods are readily tricked.
- Photo verification has been provably fooled by a variety of means, including drawing a fake beard on your face, capturing video from an a video game character, and borrowing an older friend.
- Fake IDs are readily accessible to teenagers, who already use it to buy alcohol and tobacco.
- Credit cards can be borrowed under fake pretexts, or borrowed from an older friend.
In all cases, account sharing is trivial and socially accepted as a concept. Identity verification to validate an account, rather than an individual, is fundamentally broken. Verification of an individual would require a much more invasive and persistent form of on-device surveillance, with the goal of linking much of a person’s online activity to a real identity.
That’s the point
When a system collects real-life identification, makes it hard for adult content providers and small websites to operate, and still fails to filter out minors, what you have is not an online safety system. What you have is an anti-porn censorship system and mass surveillance network.
According to POSIWID (The purpose of a system is what it does), that’s the point.
There’s a certain mindset, particular to older politicians, who fear change. They actively seek to return us to the way things used to be, to a pseudo-mythical past that resembles the idealized and easily-understood narrative about the world they believed when they themselves were younger. There is no place for Pornhub in the idealized past.
Next, the entire open, free, and anonymous Internet is anachronistic to the nostalginauts. The idea that information of any sort can be acquired discreetly, or that people can communicate without showing their face, is an anathema to their worldview. The conspiracy theorists have us running frightened of immigrants, benefits claimants, trans people, or “woke”; we should rather be afraid of the imminent and ongoing dismantling of the World Wide Web after thirty-odd years of freedom, which will strike us all regardless of political alignment.
Happy Pride month, everybody.
Further reading
- Age verification for social media – the beginning of the end for a free internet? - Mullvad. “In reality, age verification lays the foundation for a fully government controlled internet.”