Only a very lazy bad actor would try to use ChatGPT to generate propaganda (and admittedly, there's probably quite a few people who will get caught trying to do this). Anybody who's even slightly sophisticated will use a local LLM, which isn't something that you can guard against.
How OpenAI is approaching 2024 worldwide elections
31–37 of 37 posts
Re: How OpenAI is approaching 2024 worldwide elections
#32Well, no matter what happens or who you root for, this is going to be a real spectacle of a year.
Re: How OpenAI is approaching 2024 worldwide elections
#33Well, no matter what happens or who you root for, this is going to be a real spectacle of a year.
I thought as well. I cannot wait for the circus this year, is gonna be glorious. With AIG, writing BS is gonna be so much easier.
Re: How OpenAI is approaching 2024 worldwide elections
#34Earlier quoted context omitted.
A hash doesn't work without solving all of the same problems their provenance classifier solves: you need a way to be able to detect the steganogram even after the image has been modified.
> you need a way to be able to detect the steganogram even after the image has been modified. True, but not really. If you look at any suggested method to prevent the spread of CSAM "at scale" its based on large "hash databases" and not fancy classifiers (yet, if you can't develop one, and yes they have "problems" because they dont only contain hashes for "bad" files - "bad" being "locale" dependent). Also you need t…
The hashing algorithms in question require solving all the same problems as the classifiers: they're perceptual hashes not cryptographic.
Look at how NeuralHash is implemented.
Re: How OpenAI is approaching 2024 worldwide elections
#35>We are also experimenting with a provenance classifier, a new tool for detecting images generated by DALL·E "We could hash the sh*t we output but we cant be bothered, so we wrote an AI that can vaguely detect our own output - aren't we the best AI company?" Honestly, I wish they'd stop "trying to be good" and just "get on with it". "Investors" wont buy this "we're so good" line for long.
> "We could hash the sh*t we output but we cant be bothered, so we wrote an AI that can vaguely detect our own output - aren't we the best AI company?" Can you elaborate? Wouldn’t it be trivial to modify the image so it no longer has the same hash?
Since they didn't specify, perhaps they meant using perceptual hashing (because yes, modifying an image to have a different cryptographic hash is indeed trivial).
Re: How OpenAI is approaching 2024 worldwide elections
#36You don't need to use AI to create an illustration of Biden fucking a horse (or whatever), it is simply easier and faster.
And the problem isn't that people can't see what is the truth, it is not that hard to get right. People rarely care about being right, they care about winning, so they will believe things if they support their preconcieved notions.
Re: How OpenAI is approaching 2024 worldwide elections
#37Earlier quoted context omitted.
> "We could hash the sh*t we output but we cant be bothered, so we wrote an AI that can vaguely detect our own output - aren't we the best AI company?" Can you elaborate? Wouldn’t it be trivial to modify the image so it no longer has the same hash?
Since they didn't specify, perhaps they meant using perceptual hashing (because yes, modifying an image to have a different cryptographic hash is indeed trivial).
TIL. Thank you!