[HN Gopher] Subscriptions Drive Views of Extremist Videos on You...
       ___________________________________________________________________
        
       Subscriptions Drive Views of Extremist Videos on YouTube
        
       Author : alephnerd
       Score  : 11 points
       Date   : 2024-11-03 20:00 UTC (3 hours ago)
        
 (HTM) web link (home.dartmouth.edu)
 (TXT) w3m dump (home.dartmouth.edu)
        
       | cardanome wrote:
       | Effectively disabling dislikes is what has led to this mess in
       | the first place.
       | 
       | If your algorithm takes takes the like and dislike ratio into
       | account for what videos are shown then people are encouraged to
       | create content that is not offensive to the majority of people.
       | There is a natural self-policing and extreme takes get less
       | visibility.
       | 
       | If you optimize for engagement, well we all know that spreading
       | hate and general negativity generates the most engagement. So
       | content creators will focus on that kind of content. There is no
       | way for the moderate majority of the user base to keep them in
       | check.
        
         | djaychela wrote:
         | Dislikes are not disabled, just the count is not visible. You
         | can still see the ratio in your creator studio.
         | 
         | The algorithm is optimising for view time, I think that's far
         | more of a variable than Dislikes, and people will dislike
         | content but still watch a significant portion of it.
        
           | cardanome wrote:
           | Yeah, that is what I meant when I said it is optimizing for
           | engagement i.e. view time. Early youtube algo did take
           | dilikes into accound and tried to optimize for quality.
           | Dislikes are effectively disabled.
           | 
           | A high dislike count isn't something content creators need to
           | worry about, hence the more extrem content.
        
             | hhh wrote:
             | It is something they need to worry about. It's just less
             | effective for brigading videos.
             | 
             | Except now those people over-focus on the chrome extension
             | that records the like:dislike ratio of videos (of extension
             | users only) and they end up filtering themselves by
             | thinking it's authoritative for the overall view
             | population.
        
       | Syonyk wrote:
       | > _Given the challenges of trying to characterize the content of
       | every single video viewed, the researchers focused on the type of
       | YouTube channels people watched. They compiled lists of channels
       | that had been identified as alternative or extreme by journalists
       | and academics and then examined how often a participant visited
       | videos from those channels._
       | 
       | > _Alternative channels included content for men's rights
       | activists, anti-social justice warriors, and intellectual dark
       | web material; extreme channels included white supremacist, alt-
       | right, and extremist material._
       | 
       | Eh...
       | 
       | Relying on "journalists and academics disliking a channel enough
       | to claim it's extreme," but then not bothering to actually go
       | look at the details of the material watched, sounds pretty lazy
       | to me if you're trying to make this sort of argument.
       | 
       | Or to at least _sample_ a reasonable cross section (perhaps
       | videos that showed up in a lot of the monitored profiles) and
       | make sure they 're generally consistent with what the claimed
       | nature of the channel is.
        
       | 23B1 wrote:
       | The academy's fear of 'harmful' words has become more dangerous
       | than the words themselves.
        
       ___________________________________________________________________
       (page generated 2024-11-03 23:01 UTC)