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

Twitter to label deepfakes and other deceptive media

reuters.com

Post reply on HN
Print view View original

51–60 of 151 posts

Re: Twitter to label deepfakes and other deceptive media

#51

Post by blackearl »

bilekas wrote: post #48
tsukurimashou wrote: post #47

So, where do we exactly draw the line with these things? Considering everything you see on media has been altered in some ways. To me it looks like the same problem as "fake news", how do you define "fake news" considering every piece of news has at least some bias from the person that wrote it. I think it would be a better idea to educate people about these things, just like when "Photoshop" and other similar softwa…

> I think it would be a better idea to educate people about these things This argument is nonsense an innefective. Try and explain to your elders about what a deep fake is, and how to identify one. The movie : "The Irishman" used a lot of the technology successfully to take years off the cast. Applying labels to known false media is not the same as an opinion based 'fact' or alternative facts. So don't come with this…

> and other deceptive media

Ok where's the line on deceptive media? A deepfake obviously has concrete criteria to identify it. What about a little spin? Cropping out a piece of a picture or a few seconds of a video? Muting a crucial bit of audio? What about that silly scenario where a reporter goes out in a kayak to show flooded streets and a cheeky passerby walks behind them in knee-high galoshes? Those all seem deceptive. I don't trust Twitter to have my best interests at heart when deciding. Then again, I don't use Twitter so I don't really care.

Re: Twitter to label deepfakes and other deceptive media

#52

Post by bilekas »

blackearl wrote: post #51
bilekas wrote: post #48

Earlier quoted context omitted.

> I think it would be a better idea to educate people about these things This argument is nonsense an innefective. Try and explain to your elders about what a deep fake is, and how to identify one. The movie : "The Irishman" used a lot of the technology successfully to take years off the cast. Applying labels to known false media is not the same as an opinion based 'fact' or alternative facts. So don't come with this…

> and other deceptive media Ok where's the line on deceptive media? A deepfake obviously has concrete criteria to identify it. What about a little spin? Cropping out a piece of a picture or a few seconds of a video? Muting a crucial bit of audio? What about that silly scenario where a reporter goes out in a kayak to show flooded streets and a cheeky passerby walks behind them in knee-high galoshes? Those all seem dec…

> What about a little spin?

If the 'spin' consists of a lie, then how could you even consider it true ? 10% false and 90% truth does not result in a diluted truth or half truth:

'Half True' is an oxymoron. The content is take as a whole.

If I said to you : "You need glasses with a light on them because at night its always dark" - Thats spin.

If I said : "The sun is gone forever so you need glasses with a light on them" - Thats bullshit.

Use your common sense. And if its a 'grey' area then you can't call it fake can you ?

https://media.tenor.com/images/c731ffc70a71d8a89b5e3581953f9...

Re: Twitter to label deepfakes and other deceptive media

#54

Post by pequalsnp »

geofft wrote: post #41
ReptileMan wrote: post #31

Truth is too important to let the public decide what is true. I think that twitter doesn't have enough resources to bear this burden alone. So it is better if the state give a helping hand. A dedicated ministry of truth should be there when Twitter cannot hold the line against the ever more devious deceptions. Also since a lot of the speech online is hate speech we should probably also make a state organization that…

This seems reasonable, but I'm a bit worried you're being sarcastic. Are you worried about this for some reason? Maybe scared of something you read in a fiction book once? Is this a good idea in reality?

He is paraphrasing parts of 1984 by George Orwell.

Your comment must be sarcastic as well. Please tell me that you don't actually believe that the government should hold a monopoly on truth...

Re: Twitter to label deepfakes and other deceptive media

#55

Post by pjc50 »

wyldfire wrote: post #37
pjc50 wrote: post #35

There are obvious cases of deceptive media, where things are simply made up, but the most invidious ones are just "differential spin" and very hard to pin down. The recent Harry/Megan fiasco has provided a great set of examples on how the same action can be reported as a positive or negative depending on the paper's preferred spin: https://www.buzzfeednews.com/article/ellievhall/meghan-markl...

Goodness! The avocado one was two stories fewer than eighteen months apart. It almost looks like they were intending to get caught.

More that they don't care if they get caught. You could run diametrically opposing articles a week apart and not expect the readership to care enough to stop buying the paper.

Several papers run opposite stories locally: DMG Media (owners of the Mail) notoriously did this with the HPV vaccine. Irish Mail were in favour, UK Daily Mail were against, on the same day.

The Sun vs Scottish Sun is another repeat offender.

Re: Twitter to label deepfakes and other deceptive media

#57

Post by dTal »

pjc50 wrote: post #55
wyldfire wrote: post #37

Earlier quoted context omitted.

Goodness! The avocado one was two stories fewer than eighteen months apart. It almost looks like they were intending to get caught.

More that they don't care if they get caught. You could run diametrically opposing articles a week apart and not expect the readership to care enough to stop buying the paper. Several papers run opposite stories locally: DMG Media (owners of the Mail) notoriously did this with the HPV vaccine. Irish Mail were in favour, UK Daily Mail were against, on the same day . The Sun vs Scottish Sun is another repeat offender.

Sounds a bit like A/B testing with public opinion.

Re: Twitter to label deepfakes and other deceptive media

#59

Post by libertine »

irjustin wrote: post #2

I'll keep an eye on this one. I'm hopeful because we need something like this, but man it's going to be hard with LOTS of problems on both sides of the line. Guys who get labeled as fake who didn't deserve it and guys who should be labeled not getting it. All the false positive/negatives will have a terrible time.

The problem of these scalable approaches is well documented (not scientifically as far as I know), but in several forms of media.

You said well the consequences:

- False positives - the wrongfully labelled ones ;

- """hackers""" - because there's not enough amount of '"' to make it sound different but bare with me, it's just those who learn/are lucky to manipulate/circumvent the system;

Even if this are small percentage cases, it will be in the order of hundreds of thousands, if not millions of users.

How do you give support to this people? Is it going to be another Youtube fuck-fest where you end up talking to bots? Because this shouldn't be allowed - at all.

If you want to make automated control measures, then people should have the right to talk directly to a human who can actually solve their problems. Yet that seems to be impossible, because of the scale.

Re: Twitter to label deepfakes and other deceptive media

#60

Post by dTal »

bilekas wrote: post #50
pferde wrote: post #42

Earlier quoted context omitted.

If a person makes up a quote and distributes it with their own picture, is it really a made-up quote, is it "fake"? :)

"So proud to be the Man of the year 2020" - Me Its fake, just because I said it, doesn't make the content true.

This seems to be a language drift worth nipping in the bud. It's not "fake". It's a lie. When something purports to be something it's not, "fake" is an adjective describing that thing. "Lie" is the adjective that describes the claim.

If you tweet that you are Man of the Year 2020, you are fake, a phony. The tweet, on the other hand, is simply a lie.

Post reply on HN