Forscher haben mithilfe eines Multi-Task-Learning-Modells (MTL) eine neue Methode zur automatischen Erkennung von Hassreden in sozialen Medien entwickelt und dabei herausgefunden, dass rechtsgerichtete politische Persönlichkeiten Online-Hass schüren
https://www.uts.edu.au/news/tech-design/right-leaning-political-figures-fuel-online-hate
18 Comments
>A multi-task Learning model is able to perform multiple tasks at the same time and share information across datasets. In this case it was trained on eight hate speech datasets from platforms like Twitter (now X), Reddit, Gab, and the neo-Nazi forum Stormfront.
>
>The MTL model was then tested on a unique dataset of 300,000 tweets from 15 American public figures – such as former presidents, conservative politicians, far-right conspiracy theorists, media pundits, and left-leaning representatives perceived as very progressive.
>
>The analysis revealed that abusive and hate-filled tweets, often featuring misogyny and Islamophobia, primarily originate from right-leaning individuals. Specifically, out of 5299 abusive posts, 5093 were generated by right-leaning figures.
Paper: [Generalizing Hate Speech Detection Using Multi-Task Learning: A Case Study of Political Public Figures – ScienceDirect](https://www.sciencedirect.com/science/article/pii/S0885230824000731?via%3Dihub)
File that under the category of NSS.
I’m very suspicious of any sort of software which claims to be able to parse out nuances in human speech or writing with any degree of accuracy.
Research from the Left appears to find fault with the Right. Research from the Right appears to find fault with the Left. What does my bank account tell me? I had it much better during the Trump administration than the Biden puppet show. I simply cannot afford a Harris puppet theater.
Reddit can barely be bothered to remove slurs. And companies like Twitter and Meta are even worse.
Well this is unsurprising given that the right wing politicians are openly campaigning on hate speech.
Bill gates will be pleased
“Hate speech” = speech not compatible with our political agenda
Brought to you by my definition of hate, and my invalidation of groups aversion of one another (as seen demonstrated in nature)
Expect those researchers to be targeted by people like Musk, as the disinformation researchers have been.
We knew that before the study results
Everyone just needs to get off the Internet and social media.
I’m hoping AI and the further development of tools like this just push users away until no one wants to partake in a fake, censored, AI driven cesspool of garbage content that no one enjoys anymore.
Readily apparent to anyone who has been online the past 16 years. You could see it clearly in the comments sections of articles that were linked on Drudge Report, tons of right-wing, transphobic, misogynistic and generally just wild comments completely out of step with the normal conversation.
Sounds totally harmless and not anti free speech at all
I’m sure this study isn’t biased at all.
Sounds pretty dystopian to me.
Just a reminder that the congress recently defined any criticism of Israel as “anti-Semitic” and therefore hate speech.
We’re going to trust an AI run by a corporation to tell us what we are and aren’t allowed to know?
What else are they going to say? Hey everyone I know you’re struggling but so are businesses so if it’s all right with you we’re going to give them a nice tax break. Don’t worry you can all go on welfare that we designed to be painful as possible.
It’s quite funny that that they come up with the obvious, do we really need “ai” or research to verify this?