#6072494

fantaskarsef

2022-11-15 09:06

I personally think tech like this is already so important, and will become more important as time goes on. If they can really make it work like intended, every news outlet (or at least bigger news agency who do their own research) needs to have access to something like that. And if not that, than every person should have access. Now they also have to make money off of that... will see how that goes.

#6072527

k3vst3r

2022-11-15 10:29

So basically if the deepfake software used can fake a persons blood flow across the face, it can fool this deepfake detector also. Perhaps saying how it works isn't wise, because now they know what to work on to get around it.

#6072651

geogan

2022-11-15 15:30

Detecting bloodflow? I've seen enough low quality pixelated, macro-blocked mess videos to know that infomation is not contained in 99% of online video. Sounds like bull to me. And I have programmed with this image recognition OpenCV before.. Wonder what the angle is - just sell this snake-oil to very rich buyers (government) anyway and don't care if it actually works or not - think Dick from OCP strategy for ED209 in Robocop - "who cares if it works or not!". Also love they way these corporate press releases stick company name before everything - "Employees also leveraged Intel Water and Intel Food..."

#6072681

XenthorX

2022-11-15 16:15

Here's actually an initiative that could help western democracies.

#6072711

schmidtbag

2022-11-15 16:54

I find it kinda funny how they're acting like this is only possible using Intel's products. Kinda cringy how they're shoving their name in front of AVX, especially considering AMD seems less discriminatory about implementing it.

#6072739

RealNC

2022-11-15 18:12

If ML can come up with stuff that can fool humans, then you can also train it to come up with stuff that fools deep fake detectors. Just sayin'.

#6072757

Zooke

2022-11-15 18:48

96% effective eh ? That sounds familiar. I guess we can all sleep easy then knowing Intel is protecting us.

#6072871

X7007

2022-11-15 23:58

so if the program will say it's correct, would we need to agree with the software that made the deep fake software and the software made by the people to detect AI, then they would be able to be the judges to say who is fake and not? seems to me like it's a god's power to decide who is and who isn't.

#6072905

Venix

2022-11-16 02:54

RealNC:

If ML can come up with stuff that can fool humans, then you can also train it to come up with stuff that fools deep fake detectors. Just sayin'.

Another cat and mouse game begins , is what I see !

#6072949

pimpineasy

2022-11-16 08:36

↑ If ML can come up with stuff that can fool humans, then you can also train it to come up with stuff that fools deep fake detectors. Just sayin'. Another cat and mouse game begins , is what I see !

Sooo.... A deep fake making a deep fake anti fake all while a deep fake hilbert wrote the thread. Bet ooooh fooorrrr deep fake. Doomer Nation

#6074349

Andrew LB

2022-11-20 11:11

XenthorX:

Here's actually an initiative that could help western democracies.

If only it could detect fake news... that would be the end of CNN and MSDNC.