YouTube to Deploy Crappy AI to Enforce Age Verification… on Americans

American users are going to encounter an awful AI system trying to determine their age on YouTube soon.

Between the fiasco that is the deployment of AI into different services and the rapid onset of censorship through age verification that goes well beyond anything deemed to be “pornography”, it seems that YouTube has decided to combine the two bad issues into one giant mess.

First, there is the well documented many failings of AI. The problem is that corporations are just shrugging their shoulders at just how bad AI is and have collectively started saying “good enough”. This as corporations like Rogers have started replacing people’s jobs with AI. Predictably, we are already seeing those results as AI have begun deleting the very user data they were entrusted to protect and enhance.

Second, the age verification creep has been rapidly expanding well beyond websites deemed to be dealing in “pornography”. After all, these days, we are far removed from the early days of preventing minors from browsing porn websites and going well into the area of censoring anything deemed inconvenient for the government to see people consume (which was pretty predictable). We saw this as governments started demanding age verification requirements be expanded to video games, app stores, social media, and search engines.

Large web services, having left their spines at the SOPA debates, started bending at the knee to the governments censorship demands and pressure from payment processors that may or may not have anything to do with radical pro-censorship extremists. This includes Reddit, Bluesky, Roblox, XBox, itch.io, and Steam. After having conquered some of the biggest websites through cowardly surrender, we have already started seeing political speech actively being censored through the age gates getting installed. This is generally the worst case scenario I have long warned about playing out at a frighteningly fast rate.

While much of this mass government mandated censorship is being inflicted on UK residents, they are far from the only ones being affected by this pressure. Recently, we learned of one effort to bring mass government sponsored censorship to American users. From the Guardian:

YouTube announced on Tuesday that it will begin to use artificial intelligence to estimate the ages of users in the US, in order to show them age-appropriate content.

The rollout of the new feature comes one day after Australia’s government announced it would ban children under 16 from using YouTube and less than a week after the UK implemented sweeping age checks on content on social networks.

YouTube’s AI age verification on its home turf indicates it is putting into place a form of compliance with the Australian and UK requirements, despite its persistent opposition to age-check requirements.

“Over the next few weeks, we’ll begin to roll out machine learning to a small set of users in the US to estimate their age, so that teens are treated as teens and adults as adults,” wrote James Beser, director of product management for YouTube Youth, in a blogpost titled Extending our built-in protections to more teens on YouTube.

YouTube was promised an exemption from Australia’s social media ban last year by the then communications minister, but the Australian government said on Monday that the platform would, in fact, be included in the country’s ban on children under 16 using social networks. The ban is slated to take effect in December. Google, YouTube’s parent company, strenuously advocated against the Australian ban and has threatened to sue to overturn it.

Now, it goes without saying that facial recognition and AI has been notoriously bad at estimating people’s age. If you have a younger looking face or are close to the age limit, then the accuracy begins to drop off quite a bit. This research is well known, but frequently forgotten about (or ignored altogether) during these debates by large mainstream media and policy makers.

Younger creators are likely going to suffer from a rollout like this. One example is Blue Scuti who became famous for being the first player to beat Tetris into a crash. I happen to know that he has faced problems growing his audience since part of his rise to fame happened when he was also classified as “under age”. This meant that older creators had to help him pick up the slack by documenting his achievements and allow discussion for his achievements until he became old enough himself.

Blue Scuti is, of course, far from the only creator out there that rose to fame at a young age. Numerous creators started gaining traction when they were technically a minor. Because there are going to be rollouts like this, younger creators who are trying to capture a solid future at a younger age are going to be faced with additional barriers along the way, depending on how this is all deployed.

While this is a trial run for the time being, it is a sign that Alphabet might also be caving to government pressure to begin massively censoring their platforms (at least much more than they already have been in the past – looking at you ContentID!). This as we pick up momentum on this censorship slippery slope that seems to be getting worse and worse by the day.

Drew Wilson on Mastodon, Twitter and Facebook.