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

YouTube AI deletes war crimes evidence as 'extremist material'

middleeasteye.net

Post reply on HN
Print view View original

241–250 of 334 posts

Re: YouTube AI deletes war crimes evidence as 'extremist material'

#241

Post by the8472 »

djrogers wrote: post #224
the8472 wrote: post #183

Earlier quoted context omitted.

I think the problem is the expectation of people that someone else do the filtering for them. I.e. "I don't want to see this kind of content" leads to "someone else should remove it from all the sites I visit". Which obviously leads to conflicting requirements once you have more than one person and those people disagree on what they want to see and don't want to see. The only reasonable solution is to host everything…

> I.e. "I don't want to see this kind of content" leads to "someone else should remove it from all the sites I visit". I don’t think that’s always the mindset. Isn’t it reasonable for people to have the mindset of “I want to go to sites that don’t have content that I find objectionable”? This way websites can decide which group of people they want to cater to.

The context of the discussion is large websites acting as platforms. Their users are bound to have conflicting views about what's "objectionable". So when the moderation mechanism is deletion instead of letting users just filter then the website has to preferentially treat one group instead of being a platform for everyone.

Re: YouTube AI deletes war crimes evidence as 'extremist material'

#242

Post by dillweed »

viridian wrote: post #29
Hasknewbie wrote: post #9

Youtube's response regarding one of these videos documenting abuses (emphasis mine): > "we've determined that your video does violate our Community Guidelines and have upheld our original decision. We appreciate your understanding ." Can someone explain to me why corporations, when interacting with customers regarding complaints/appeals, seem to have "don't forget to add insult to injury" as one of their motto more o…

If I remember correctly, the exact same message was delivered to Jordan Peterson a couple of weeks ago or so, before he sat down with the google memo guy. He was in the middle of a bible lecture series, and Google banned his account, and sent the exact same "we've determined that your video does violate our Community Guidelines and have upheld our original decision. We appreciate your understanding." message. It seem…

It's a strange corporate culture thats for sure

Re: YouTube AI deletes war crimes evidence as 'extremist material'

#243

Post by mirimir »

vidarh wrote: post #185
the8472 wrote: post #183

Earlier quoted context omitted.

I think the problem is the expectation of people that someone else do the filtering for them. I.e. "I don't want to see this kind of content" leads to "someone else should remove it from all the sites I visit". Which obviously leads to conflicting requirements once you have more than one person and those people disagree on what they want to see and don't want to see. The only reasonable solution is to host everything…

The problem on these distributed platforms is not filtering what people see , but filtering what people host or allow to transit their network connections. > In a decentralized system you also skip the law requirements since you cannot enforce multiple incompatible jurisdictions at the platform level, individual users will be responsible for enforcing it on their own nodes, similar how all you can do when accidentall…

> But that's the thing: You don't skip it. You spread it to every user. They both have to deal with whether or not they are willing to host the material and whether or not it is even legal for them.

> How many of us sympathise with the idea of running a Tor exit node, for example, but avoid it because we're worried about the consequences?

Tor isn't the best example, because exits don't cache anything. So mainly, exit operators get complaints. And the exit IPs end up on block lists. Operators don't typically get prosecuted. Maybe they get raided, however, so it's prudent to run exit relays on hosted servers.

Freenet is the better example. The basic design has nodes relaying stuff for other nodes. In an extremely obscure and randomized way. Also, keys are needed to access stored material.

However, nodes see IPs of all their peers. Investigators have used modified clients to identify nodes that handle illegal material. So users get busted. There is "plausible deniability". But it's not so plausible when prosecutors have experts that bullshit juries. So users typically plea bargain. Or, if they use encrypted storage, they get pressed for passwords. Like that guy in Philadelphia.

Re: YouTube AI deletes war crimes evidence as 'extremist material'

#244

Post by redthrowaway »

One of the most interesting developments in AI will be watching how we respond to human rationality detached from human morality. Programs that optimize for practical outcomes are going to come up with a whole host of solutions that we consider abhorrent, not least because the mere notion that that solution is a practical one riles our sensibilities.

Re: YouTube AI deletes war crimes evidence as 'extremist material'

#245

Post by AmIFirstToThink »

Buge wrote: post #239
AmIFirstToThink wrote: post #199

Why not create a setting that allows user to see YouTube as sanitized by their AI or all content? Allow people to chose content level just like they choose security level in browser settings. 1. Legal content. May include content that violates YouTube content policy, but is legal in USA, or the country of the viewer. Maximum freedom of speech and maximum ability to see content that you may find offending. 2. YouTube…

The whole point of the censorship is to stop Youtube from being a tool used for terrorist recruitment. If the terrorists can just check the box "all content" then it's useless.

Terrorist induction videos would clearly fall into Not-legal category. I think NSA can easily stand up resources to scan new videos as they arrive and tag as legal or not-legal. Or, we can demand a judge's decision on taking away someone's freedom of speech as the constitution clearly indicates.

And, I have a hunch that three letter acronym gov organizations dealing with keeping us safe would rather know who is interested in terrorist induction videos, and track them, than just remove the videos and let opinions fester at an isolated individual level (lone wolf).

Re: YouTube AI deletes war crimes evidence as 'extremist material'

#246

Post by AmIFirstToThink »

nthcolumn wrote: post #210
AmIFirstToThink wrote: post #155

What did they train the AI on to deem something 'extremist'? Should we get to see the training data used and labels? Or is this the modern day equivalent of credit score algo, something that can have huge impact on lives, but you are not allowed to know what it is. This is bad.

We will have to get used to this - people hiding behind their AI's skirts saying 'Wasn't me - she did it'.

They are still liable. They unleashed it upon the world.

Can't wait till they are sued out of their bubble.

Re: YouTube AI deletes war crimes evidence as 'extremist material'

#248

Post by AmIFirstToThink »

colejohnson66 wrote: post #208
AmIFirstToThink wrote: post #199

Why not create a setting that allows user to see YouTube as sanitized by their AI or all content? Allow people to chose content level just like they choose security level in browser settings. 1. Legal content. May include content that violates YouTube content policy, but is legal in USA, or the country of the viewer. Maximum freedom of speech and maximum ability to see content that you may find offending. 2. YouTube…

Because YouTube doesn’t care about the videos; it cares about the advertisers. You can’t be a proponent of free speech (extremist propaganda) while trying to please advertisers. Also, with today’s political climate, people seem like they want anything that disagrees with them to be labeled as hate speech.

The advertisers would find it suitable to chase a bubble e.g. greens, nationalists, globalist, feminist, religions etc.

I think a good strong case can be made to advertisers that their ad will only be served to people opting into a certain bubble. Or reverse of that i.e. show my Ads to all people except those who are in this bubble. Inclusion list and exclusion list.

Re: YouTube AI deletes war crimes evidence as 'extremist material'

#249

Post by AmIFirstToThink »

cvsh wrote: post #32
viridian wrote: post #29

Earlier quoted context omitted.

If I remember correctly, the exact same message was delivered to Jordan Peterson a couple of weeks ago or so, before he sat down with the google memo guy. He was in the middle of a bible lecture series, and Google banned his account, and sent the exact same "we've determined that your video does violate our Community Guidelines and have upheld our original decision. We appreciate your understanding." message. It seem…

The worst part of the information age is arbitration by unreasonable and impenetrable algorithms rather than humans with the capacity to make a judgement call when the rules clearly don't account for the situation at hand. A transparent appeals process staffed by humans who can at least deliver a rationale, including what rule you broke, should be required by law. There's irreparable reputational damage associated wi…

I mean, why not wait till it is deemed illegal by the authorities and due process? How can you take away someone's freedom of speech?

It's not like the AI has absolute idea of what is 'extremist' content, it's just enforcing someone idea of what it is. AI is trained on data, and whoever labeled that data is the person/s who are winning here.

Re: YouTube AI deletes war crimes evidence as 'extremist material'

#250

Post by LoSboccacc »

giancarlostoro wrote: post #45
jacobr wrote: post #2

Also see this Twitter thread: https://twitter.com/EliotHiggins/status/896358097320636416 > Ironically, by deleting years old opposition channels YouTube is doing more damage to Syrian history than ISIS could ever hope to achieve > Also gone are the dozens of playlists of videos from Syria I created, including dozens of chemical attacks in playlists by date > Keep in mind in many cases these are the only copies of the…

That is a little insane... Makes me wish there was some sort of website that archived specific YouTube videos marked as historical or criminal evidence or some sort of qualification, as long as they don't abuse copyright just to make sure if places like YouTube delete them they can remain. Or a service that uploads to multiple video streaming sources at once (though I imagine these might violate the TOS of YouTube fo…

> Makes me wish there was some sort of website

Can't people just use liveleaks for that kind of videos?

Post reply on HN