Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Stopping deepfake news with an AI algorithm that can tell when a face doesnt fit

spie.org

Post reply on HN
Print view View original

1–10 of 105 posts

Re: Stopping deepfake news with an AI algorithm that can tell when a face doesnt fit

#2

Post by turblety »

Now we can use this new AI to train another AI that can defeat it. And so continues the great cat vs mouse chase.

I don't think this is a problem that is ever going to be solved. Deep fakes will become more and more popular, and harder to detect.

Re: Stopping deepfake news with an AI algorithm that can tell when a face doesnt fit

#6

Post by giancarlostoro »

DarthGhandi wrote: post #4
formerly_proven wrote: post #3

Isn't this the idea of getting better results using the adversarial network approach? So this would be inherently ill-suited to stopping deepfake news, yeah?

it's great at making them better, yes

If the people trying to stop deepfakes can do it so can the people trying to produce them anyway. The best we can do is some way of digitally signing videos and software displays it as authentic or not but then they just make fake youtube style sites that say its authentic. The problem will go on and on.... Fake news will always spread for one reason or the next.

Re: Stopping deepfake news with an AI algorithm that can tell when a face doesnt fit

#7

Post by IgorPartola »

turblety wrote: post #2

Now we can use this new AI to train another AI that can defeat it. And so continues the great cat vs mouse chase. I don't think this is a problem that is ever going to be solved. Deep fakes will become more and more popular, and harder to detect.

I feel like it’s time to bring in Heinlein’s Fair Witness concept: https://www.urbandictionary.com/define.php?term=Fair%20Witne...

If you think about it, when we had relatively few journalists and their reputation was on the line with every story they wrote they in a way were fair witnesses. Since we have moved to a hybrid model of professional journalists and a huge caste of YouTubers, podcasters, bloggers, and twitters, it has become impossible to hold everyone to a high journalistic standard. At the same time Fox News has brought on an era of politicized news so the bias is now inherent. Maybe if we had a few hundred fair witnesses that offered a professional service we could have a way to verify facts. Then again, it’s one thing to have a fair witness present facts, it’s another to have that witness’s likeness be deep faked and broadcast all over the web saying something they didn’t actually say.

Another solution may be regulatory: anyone broadcasting a deep fake with a huge disclaimer that it is a fake gets a 20% of their net worth fine. Draconian but I can see it working.

Re: Stopping deepfake news with an AI algorithm that can tell when a face doesnt fit

#10

Post by ben_w »

IgorPartola wrote: post #7
turblety wrote: post #2

Now we can use this new AI to train another AI that can defeat it. And so continues the great cat vs mouse chase. I don't think this is a problem that is ever going to be solved. Deep fakes will become more and more popular, and harder to detect.

I feel like it’s time to bring in Heinlein’s Fair Witness concept: https://www.urbandictionary.com/define.php?term=Fair%20Witne... If you think about it, when we had relatively few journalists and their reputation was on the line with every story they wrote they in a way were fair witnesses. Since we have moved to a hybrid model of professional journalists and a huge caste of YouTubers, podcasters, bloggers, and twit…

We need something, but Fair Witnesses seem both exploitable and unachievable given the way human memories work.
Post reply on HN