A moderation system that leans on automation just knocked legitimate tech tutorials and even entire channels offline. The appeals felt automated, too. Creators are powerless against opaque enforcement and the incentives that should favor craft and trust are tilting toward noise.

Over the past four years, I’ve significantly reduced my social media footprint. There are countless reasons for this, all of which are beyond the scope of this article, but the point I want to make is this: despite my growing apathy and downright hostility towards social platforms, I’ve found YouTube to be an oasis of sorts.

I am not going to pretend that YouTube hasn’t played its part in the global disinformation epidemic or that it has somehow escaped the claws of enshittification. What I will say is that unlike other social platforms, its feed (unlike those of its competitors) are maleable using browser-based plugins (tools such as subscription managers). It is one of my primary learning platforms; without its vast array of tutorials, there is no way that I, a non-programmer, would have learnt Linux as fast or become as comfortable in a FOSS-based computing environment, as I have since the pandemic.

But enshittification is, like death and taxes, a certainty now. Which brings us to the subject of this column: AI moderation on YouTube.