AI Slop has finally met its match Google Research Paper: https://research.google/pubs/scalable-detection-of-adversarial-synthetic-slop-and-coordinated-media-abuse-a-lora-enabled-multimodal-defense-system/ Slop Purge Article: https://insidethecreator.beehiiv.com/p/the-slop-purge-has-arrived Pokemon Channel ► https://www.youtube.com/@dolandarkrai Main Channel ► https://www.youtube.com/c/dolandarker Second Channel ► https://www.youtube.com/c/dolandark
ADVERTISEMENT
"A broken clock is right twice a day." -Sun Tzu or something, idk
"google just took down 100000 ai slop channels" ... sentence got cut off ... "that didnt belong to them" now it makes sense
The fact that they actually use the term “AI slop” is killing me coming from YouTube
YouTube does 1 good thing every so often to make the next 5 bad things they're about to do look less bad.
The "not targeting legitimate AI artist channels" kinda makes me think they'll keep the giant children's AI brainrot channels and just ban the less profitable AI channels.
Whenever someone is describing themselves as an AI-artist, I already know I'm talking to a lower lifeform
Im too stockholm'd to hear good news. It's supposed to only get... a whole lot worse
They should get rid of porn bots first
2:02 “legitimate AI artists” being said in a research paper about ai slop is one of the best examples of irony
"AI being used to fight AI slop and accidentally proved AI is the only thing that can stop AI" What in the Terminator are we living in now?
The slop has a 92-95% accuracy rate
Despite everything Literally everything You’re still dark, Dolan dark
Google tech boys: " Look at this! We can identify and remove AI generated content! Google marketing team: "Really? That's great! Let me see the paper." _<puts paper in shredder>_
I feel sorry for those 8% legitimate content creators who will get their channel removed by the AI and will have no way to reach Youtube as usual.
I feel like the "we don't want to hurt the AI artists" part mean nothing will change
1:20 The 92-95% precision means that 8 to 5% of decisions made by the AI were either a false positive or a false negative. Since they write that they prioritize minimizing false positives I would assume a majority of those imprecise decisions made by the AI system are false negatives. Later in the paper they actually write that the overturn rate is below 1%, but an actual number. If it is close 1% it is still a lot of overturns but it could also be closer to .001% so we don't really know.
Is today Opposite Day, how unheard of that YouTube would actually do something good.
and yet they just added an "ask AI" tab on the right of the desktop version of youtube
people are gonna say "youtube is healing" and then youtube is gonna ban long form videos or smth lol
"Legitimate AI artists" sounds like a misnomer.