Hacker News new | past | comments | ask | show | jobs | submit login

I'm probably wrong here but don't you need a ton of real underlying content to make a deepfake? Like the reason you only see convincing ones of world leaders and celebrities is that they have their face out there a ton.

By that logic you would need a ton of actual CP to create CP deepfakes.




I mean more generating like artificial humans: https://boredhumans.com/faces.php

Some some tech like the above. Maybe the term deep-fake is not the correct one


Exactly. You need a lot of data to train a GAN.


Hmm good point. But what about honey-pots? That seems like a very easy thing to do as they already have the dataset, no?

Plus you only need images of children in general - not even CP to create those 3 models and then it's just Unreal engine let's say.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: