People are worried that AI is /will make it too easy to fake audio and video. Wondering: Could there also be AI tools to detect and sequester fakes. Kind of like what happened with spam mail?
-
Show this thread
-
Replying to @B_resnick
The whole field of GANs: generators make fakes and discriminators try to detect them. They train each other
1 reply 0 retweets 0 likes -
Replying to @jbohnslav
o... so building detectors would be an unending arms race? Basically wondering if there are potential tools (like a spam filter of sorts) to quickly identify false images / videos
1 reply 0 retweets 0 likes -
Replying to @B_resnick
https://en.wikipedia.org/wiki/Generative_adversarial_network … this is often how fakes are generated, via this arms race. I'm not sure of any work that has used human labels to explicitly try to detect GAN-generated images
1 reply 0 retweets 1 like
Replying to @jbohnslav
thanks! I'm going to read up on this.
1:07 PM - 7 Feb 2018
0 replies
0 retweets
0 likes
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.