Why is youtube suggesting aggressive actions against protestors?
cross-posted from: https://lemmy.world/post/2044512
I made sure to remove cookies and not sign in so I think these are the base suggestions made by youtube.
You're viewing a single thread.
Google is currently defending itself in the US Supreme Court over a lawsuit that alleges they assisted the terrorist group ISIS in recruiting members after it was found the YouTube algorithm promoted ISIS recruiting videos to young men who later committed a terrorist attack.
So to answer your question using Google's argument: they have so many videos that an advanced search feature is required to make the site usable. Their search feature only suggests things that are popular. It's not their fault ISIS recruitment (or other violent content) videos are popular.
The counter argument is: Google is curating content by displaying things people didn't search out themselves. This is direct promotion by Google itself and therefor it should be treated as if they are the publisher of that content. Anyone publishing violent content should be held liable for it.
Promoting terrorism increases engagement
Yeah, except it gets more complicated than that. Google wouldn't necessarily be promoting it either. As their algorithm looks for popular searches. Terrorism seems to be an overly used word for comparing protests to terrorism.
As an example, I live in a pretty red state. I would consider my self democrat/liberal in this state. When the George Floyd protests were happening a lot of people in my state were referring to the non-protest raids as terrorism. Despite the fact they will all defend the very clear terrorism on the capital as an attempt to save the U.S.
Point being you take the word protest and terrorism. You set it side by side as an exaggeration for literally anything half the the country disagrees with and boom you get popular terrorists searches.
I also don't think Google isn't at fault since their algorithm is designed to continue feeding that kind of content and the deeper you go the more ingrained into content you get and the more insane it gets.
So wait, is Google only suggesting things that are popular, or are they displaying things people didn't search out themselves? How do they prove that in court, do they need to show their source code or something?
I'm not a lawyer and haven't read through the court documents, but from legal commentators it seems that Google provides the general steps for how their algorithms work in plain language for the judges to consider. Even then, the Supreme Court itself has stated they have no clue how technology works so this is difficult for them to rule on
What prevents Google from lying about how their algorithms work, though? How could it actually be verified? There's no way it could just be as simple as they give their word and suddenly that's good enough for a court ruling?