[challenger22]

I imagine there's boiler room full of guys somewhere tagging videos as "Borderline content", creating data which then feeds into a machine learning algorithm which decides which videos get promoted. All videos are metricized by the similarity to borderline content, and are increasingly likely to be promoted in proportion to increasing graph distance from the borderline content. Youtube management gets nice plausible deniability that they don't actively surpress political adversaries, because the whole process is stochastic, and who knows how machine learning works anyway?

[cagenut]

the machines won't take over because someone puts them in charge, they'll take over because everyone's ass is covered avoiding decisions.

[googlethrosdjn]

I have personal experience with working on recommendations at google and you are pretty off. It was based on increasing engagement, user satisfaction, quality and on reducing controversial news coverage of google(this meant removing the right-wing content which could lead to media outrage no more no less). In addition there were a lot of left leaning googlers who had a personal agenda to suppress all right-leaning content.

You are right that it was a room full of guys, but what you may not realize is a majority was Chinese or Indian.