SoylentNews is people
SoylentNews
AI Trained on 4Chan Becomes 'Hate Speech Machine':
AI researcher and YouTuber Yannic Kilcher trained an AI using 3.3 million threads from 4chan's infamously toxic Politically Incorrect /pol/ board. He then unleashed the bot back onto 4chan with predictable results—the AI was just as vile as the posts it was trained on, spouting racial slurs and engaging with antisemitic threads. After Kilcher posted his video and a copy of the program to Hugging Face, a kind of GitHub for AI, ethicists and researchers in the AI field expressed concern.
The bot, which Kilcher called GPT-4chan, "the most horrible model on the internet"—a reference to GPT-3, a language model developed by Open AI that uses deep learning to produce text—was shockingly effective and replicated the tone and feel of 4chan posts. "The model was good in a terrible sense," Klicher said in a video about the project. "It perfectly encapsulated the mix of offensiveness, nihilism, trolling, and deep distrust of any information whatsoever that permeates most posts on /pol."
According to Kilcher's video, he activated nine instances of the bot and allowed them to post for 24 hours on /pol/. In that time, the bots posted around 15,000 times. This was "more than 10 percent of all posts made on the politically incorrect board that day," Kilcher said in his video about the project.
AI researchers viewed Kilcher's video as more than just a YouTube prank. For them, it was an unethical experiment using AI. "This experiment would never pass a human research #ethics board," Lauren Oakden-Rayner, the director of Research at the NeuroRehab Allied Health Network in Australia, said in a Twitter thread.
"Open science and software are wonderful principles but must be balanced against potential harm," she said. "Medical research has a strong ethics culture because we have an awful history of causing harm to people, usually from disempowered groups...he performed human experiments without informing users, without consent or oversight. This breaches every principle of human research ethics."
Just because something can be done doesn't mean it should be done. What are you views? Is this a harmless prank, a justified experiment or something potentially more sinister?
(Score: 3, Insightful) by stretch611 on Thursday June 09 2022, @06:48AM (13 children)
So an AI trolls the human cesspool known as 4chan. Does anyone even care if the human garbage there gets insulted? (and tbh, people there should expect this when they post.)
The only small problem is all the electricity wasted on the AI to be there. And lets face it, the server the website sits on is just as wasteful.
The real benefit of an AI on that board... at least the AI can be disconnected and turned off... if only that was the case for the other posters the internet would be a better place.
Now with 5 covid vaccine shots/boosters altering my DNA :P
(Score: 1, Touché) by Anonymous Coward on Thursday June 09 2022, @07:55AM (8 children)
Such a pesky bother you can’t send us wrongthinkers into the gas chambers yet.
(Score: 4, Informative) by DeathMonkey on Thursday June 09 2022, @03:58PM (7 children)
We would take your concerns more seriously if you would stop taking swastika flags to your protests.
(Score: 2, Insightful) by Anonymous Coward on Thursday June 09 2022, @05:45PM (6 children)
Didja see the video of the Republican senator whining that Rs suffer so much political persecution they can't even lie to the FBI or Congress without getting in trouble!! The comment would have floored me if I hadn't been prepped by the years of R idiocy.
(Score: 4, Interesting) by DeathMonkey on Thursday June 09 2022, @05:53PM (1 child)
Haha yeah that's a good one!
I like this one about how half the GOP election deniers who were running for governor got kicked off the ballot for fraudulent siugnatures. [michiganadvance.com] And then this week one more gets arrested for (presumably) actually trying to overthrow the last election! [apnews.com]
Just a reminder that all Republican accusations are actually confessions.
(Score: 0) by Anonymous Coward on Friday June 10 2022, @06:31PM
> Just a reminder that all Republican accusations are actually confessions.
Yep, classic projection. They're doing it so obviously everyone else must be doing it. Uh, no.
(Score: 2) by DeathMonkey on Thursday June 09 2022, @06:08PM (3 children)
Oh haha and a new contender has entered the ring!
GOP Candidate Carl Paladino Calls Adolf Hitler 'The Kind Of Leader We Need Today' [huffpost.com]
(Score: 2) by cmdrklarg on Thursday June 09 2022, @06:48PM (2 children)
Heh, perhaps his intentions with the full statement were taken out of context as Stefanik claims, but he should know better than to offer praise like that about Adolf Hitler. *smh*
Answer now is don't give in; aim for a new tomorrow.
(Score: 0) by Anonymous Coward on Friday June 10 2022, @07:44PM (1 child)
America should have never entered the war, unless it was to help Hitler. Hitler was a great racial hero. HH!
https://odysee.com/TGSNTtvPart01:eaa173ceb94bb7db302c30b826bf590a1ca49f26?lid=07d0d85ca4cbb4b08fde00850f240d5d7225795c [odysee.com]
(Score: 0) by Anonymous Coward on Friday June 10 2022, @09:25PM
There is one SN nazi outing themselves. Still AC, guess they're waiting till their orange fatass takes over everything before these knuckle draggers will feeeeel safe enough to show everyone their true side.
(Score: 2) by HiThere on Thursday June 09 2022, @01:15PM
Actually, there may be a theoretic benefit (i.e., a benefit in the development of theory). This is a test of the assumption that part of human speech patterns can be readily modeled. If, of course, it really couldn't be distinguished from a human poster.
I tend to think it's' true, but I can't imagine reading the garbage that must have been produced to validate that, and it would take more than just reading. You'd need at least a double blind study with lots of readers trying to make the distinction.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 3, Insightful) by DeathMonkey on Thursday June 09 2022, @04:03PM (2 children)
It wasn't "trying" to troll. It was trying to make good posts based on the forum content it was fed. And the "good" posts it produced based on that input was racist bullshit.
What that says about the content of the forum is left as an exercise for the reader.
(Score: 1) by khallow on Friday June 10 2022, @05:45PM (1 child)
(Score: 1, Informative) by Anonymous Coward on Friday June 10 2022, @07:46PM
I think it was ultimately meant to generate YouTube views.