Re:DALL·E Now Available Without Waitlist
by marshray on Friday September 30, 2022 @03:57PM
(#33039978)
If you develop technology that can reliably detect deepfakes (i.e., fn is_deepfake(m: media) -> bool), then couldn't an AI be trained to defeat it?
[ Parent
| on HN ]
Re:DALL·E Now Available Without Waitlist
by waste_monk on Tuesday October 04, 2022 @12:34AM
(#33076731)
Yes... but you can then train a model to detect the improved fakes. And then train a better deepfake model, and so on, forever.
This technique of pitting two AI against each other (a generator and a detector/discriminator) is called a generative adversarial network, and it's used a lot for unsupervised training.