twitter / the-algorithm

Source code for Twitter's Recommendation Algorithm
https://blog.twitter.com/engineering/en_us/topics/open-source/2023/twitter-recommendation-algorithm
GNU Affero General Public License v3.0
61.7k stars 12.17k forks source link

Suggestion for: Twitter promoting misinformation #1586

Open sagehawk opened 1 year ago

sagehawk commented 1 year ago

Twitter is promoting misinformation. image Twitter is the most susceptible to misinformation because not everyone knows blocking or reporting will debuff a tweet, instead they engage and unintentionally promote the tweet. Look at #538 where I thoroughly review the issue and suggest a solution.

ghost commented 1 year ago

Hold up. Is this true? There's misinformation on Twitter?

TolstoyDotCom commented 1 year ago

Biden, Trump, Xi, Putin, Stalin, Hitler, Mao... all have/had their own definition of "misinformation". Whose should we use? Should we trust Vijaya or Musk to determine what we can see and talk about? Or, should we pull our socks up, act like grown-ups, and engage those we disagree with and try to show them wrong? I suggest the latter.

sagehawk commented 1 year ago

@TolstoyDotCom you're absolutely right about misinformation being subjective, and I completely agree with you.

The real issue is moreso in Twitter being number 1 at amplifying objectionable content. Twitter's algorithm promotes objectionable tweets because they receive the most engagement. Whether the engagement be negative or criticism, users that respond to a tweet inadvertently promote the tweet.

I opened issue #538 for further discussion towards a solution. I welcome criticism to be certain if this is the appropriate solution.

Thank you for your feedback.

TolstoyDotCom commented 1 year ago

Give me an example of an "objectionable tweet". I mean one based on opinion, not NSFW etc.

I've seen lots of things I object to and I can always show them wrong to the target audience. Twitter makes that difficult. For instance, Leader A might say "the sky is green". I'll reply, "no, it's blue [scientific explanation]". Then, the fans of Leader A will downvote (if given that chance by Twitter), block, and/or report. That will then make it less likely that others will see how Leader A is wrong.

For those who think this is a partisan issue, a group of conservatives tried to hide content they didn't like on Digg several years ago: https://medium.com/civic-media-project/bury-until-they-change-their-ways-the-digg-patriots-as-user-generated-censorship-74dddf381a12

Clouud9 commented 1 year ago

I do like the idea, but I'm a bit afraid about this being manipulated by groups to downrank tweets of people they don't like. It might be better if it was only tailor made for the user, but that wouldn't necessarily help reduce exposure for other people. The same problem would be there if it only downranked the tweet for similar accounts/mutuals as well, since it could just reinforce an echochamber.

I think there needs to be more discussion on how this would be implemented to see if it truly a good and workable idea, and how it would be implemented if so.

imannms000 commented 1 year ago

Twitter is promoting misinformation. image Twitter is the most susceptible to misinformation because not everyone knows blocking or reporting will debuff a tweet, instead they engage and unintentionally promote the tweet. Look at #538 where I thoroughly review the issue and suggest a solution.

Where and when did you get this data? If this is legit, fb and youtube should be higher than twitter.