Comment

Comment on Teens exploited by fake nudes illustrate threat of unregulated AI

Photoshop was always something that required skill, and a computer to run it, and a copy of a paid program.

This stuff does not need a lot of those hurdles. It’s all about ease and how it’s usable on your pocket computer that you and all your classmates have with you all the time.

Your point is still fair. But there are big differences between what was and this new stuff. You woulda needed a ton more skill and the alignment of a bunch of things to casually generate fake nudes like the ones covered by this article.

source

Sort:hotnew top

RobotToaster@mander.xyz ⁨1⁩ ⁨year⁩ ago

and a copy of a paid program.

I pirated photoshop when I was 13

source
- paraphrand@lemmy.world ⁨1⁩ ⁨year⁩ ago
  Sweet. Me too. This is partially where I draw my understanding of the situation from.
  
  Specifically:
  
  The idea that piracy of professional software isn’t as casual as phone apps or web apps.
  
  The fact that it’s paid software because it’s professional software with a learning curve.
  
  source
- Deceptichum@kbin.social ⁨1⁩ ⁨year⁩ ago
  What is warez?
  
  source
Deceptichum@kbin.social ⁨1⁩ ⁨year⁩ ago
You need a hell of a lot more computer power to run a llm than you do photoshop.

source
- JohnEdwa@sopuli.xyz ⁨1⁩ ⁨year⁩ ago
  While true, you can generate AI images with a potato, it just takes longer. For my setup, RTX 3060 generates the basic image in around 10 seconds while running on CPU only takes around five minutes, but the result is exactly the same.
  
  source
  - rar@discuss.online ⁨1⁩ ⁨year⁩ ago
    Do the files exactly match to their hashes? I wonder if there’s a fundamental difference generated by using different hardwares.
    
    source
    echodot@feddit.uk ⁨1⁩ ⁨year⁩ ago
    AI always generate different outputs for the same input (AI appears to be non-deterministic) so it would be impossible to confirm that exactly.
    
    But I suppose what they mean is they appear to be of the same quality. Taking a longer time does not appear to decrease the quality of the output.
    
    I suppose you could give an AI the same input resetting it after each input and then use statistical models to identify common traits. Then do the same thing on different hardware and run the same statistical analysis and see if there is a difference between group A in group B but as far as I’m aware no one has done this.
    
    In theory hardware shouldn’t matter, it’s all mathematics basically and one plus one is always equal two, so there shouldn’t be any fluctuations.
    
    source
    -> View More Comments
- paraphrand@lemmy.world ⁨1⁩ ⁨year⁩ ago
  I assume these models are being run on servers.
  
  source
  - theterrasque@infosec.pub ⁨1⁩ ⁨year⁩ ago
    Which server run model allows pornography of this type?
    
    source