News for hackers. Stuff that matters.
Re:A.I. Muddies Israel-Hamas War in Unexpected Way by jacknews on Saturday October 28, 2023 @11:07AM (#38050367)
There's nothing unexpected about this, and the clue is in the title, it's an old technique called 'mudying the waters'.

When people can't trust some of the news, it erodes their trust in all of it, and then they can't know for sure what's actually going on.

But why bother with such subtle techniques when you can just cut communications entirely on the eve of your assualt, and then act with impunity.

Re:A.I. Muddies Israel-Hamas War in Unexpected Way by xg15 on Saturday October 28, 2023 @11:41AM (#38050672)
This. Even the current variant with generative AI in the mix isn't unexpected since the moment Musk tried to argue that basically video evidence as a category should be disallowed as, in theory, any video everywhere could be fake.

Thankfully he got laughed out of the courtroom for that. Yet.

https://www.npr.org/2023/05/08/1174132413/people-are-trying-...

Re:A.I. Muddies Israel-Hamas War in Unexpected Way by AnthonyMouse on Saturday October 28, 2023 @12:13PM (#38050945)
Elon Musk exaggerates because he's Elon Musk, but it is generally a problem.

If you have a clip of someone saying something in a news story from 2003, on the website of a news organization, with a publication date in 2003, to claim it's a deepfake would require you to claim the news organization forged the publication date. Every news organization that contemporaneously published the same video.

But if we don't have reliable means to identify deepfakes (the high quality ones, not just the bad ones), well, sometimes the video in question wasn't widely published in a year before deepfakes were a thing. So what now? If you can't tell the difference in a large number of cases then your choices are admitting the forgeries into evidence or excluding the legitimate videos.