[lotsofpulp]

That is not an “algorithm” unique to YouTube. See 24/7 news channels for a much earlier example. It is simply the nature of loosening standards on broadly available media, and throughout history, even strict standards have not always prevented the “bad” stuff from getting through.

[pavlov]

News channels don’t show random 30-minute programs created by viewers themselves. YouTube does.

Fox News and CNN may have low journalistic standards, but at least they have some. They also have liability. (Fox paid $787 million to a voting equipment manufacturer as settlement for lies they published in relation to the 2020 election.)

YouTube has neither. Their algorithm will happily promote any nonsense that has traction. The lies that cost Fox $787 million continue to circulate on YouTube unabated — and an untold number of other lies too. Alphabet has no reason to prevent this.

[ethbr1]

The greatest sin of YouTube's current recommendation algorithm is its optimization for eyeball time (aka more ad capacity).

Any tweaks around the edges will never be able to compete with that.

And unfortunately that central tenet incentivizes creators to make clickbait content that plays on emotions, because that's the most reliable way to deliver what YouTube wants.

(YouTube could decide it was optimizing for something else, but that would put a big dent in ad revenue)

[CamperBob2]

How do you fix this without doing something even worse?