YouTube blocked chess channel after mistaken detection of racial slurs • Hacker News Board

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#251

Post by amirkdv » Mon, Feb 22, 2021, 3:16 AM UTC

Aerroon wrote: post #126

das4119 wrote: post #85

Earlier quoted context omitted.

Frankly the whole thing stinks. On the whim of an unaccountable AI basically anything can happen. We need laws around this sooner rather than later. AIs are getting more powerful but less responsible. Just don't have an argument about a book or you could end up on death row. Gordon Dickson wrote a short story about this https://en.wikipedia.org/wiki/Computers_Don%27t_Argue

Why do we need laws around this? We need to hold the people accountable that put this into production and/or produced it. People should be able to build bad AI, but if that causes poor real world outcomes, then the people that okay'd its use need to take responsibility. "An AI erroneously banned the channel," shouldn't get any more leeway than "A support staff member banned the channel."

> Why do we need laws around this? We need to hold the people accountable [...] [T]he people that okay'd its use need to take responsibility.

Genuinely asking: what do you think profitable companies would do to their ML engineers over a product issue like this if there was no law under which they would be sued?

I think accountability without an enforcement mechanism is just a wishful suggestion.

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#252

Post by lovecg » Mon, Feb 22, 2021, 3:17 AM UTC

lmilcin wrote: post #74

It gets more ridiculous with every day. I fully expect people to start developing vocabularies to harden their group communication against these automatons. Also, please, don't use term "AI". Calling it intelligence is too charitable for something that can't distinguish game of chess from hate speech.

Sort of like Soviet dissident authors developed a roundabout way of writing to bypass the “algorithm” of the state’s censorship bureaucracy.

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#253

Post by joker3393 » Mon, Feb 22, 2021, 3:17 AM UTC

waterhouse wrote: post #30

haunter wrote: post #21

inb4 someone will suggest that FIDE should change the color names to be more inclusive + why does white have the first move advantage, why can't black be the first. Perfect breeding soil for just another day of pointless race drama on Twitter

You joke (I think), but given what happened to the "master" branch on GitHub, it seems likely that some people will try to push the point, and, if they make enough noise, it's likely that the organizations involved will eventually cave (not the chess grandmasters (!), but the administrator types). I googled a bit and turned up a couple of news articles: https://apnews.com/article/184386cc09be6c69535919402a13fa3f and…

"Master", as in "Grandmaster" is obviously problematic, with its obvious reference to north-American slavery.

They should change to Grand Wizard or something like that.

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#254

Post by exporectomy » Mon, Feb 22, 2021, 3:17 AM UTC

vvanders wrote: post #139

ben509 wrote: post #50

Earlier quoted context omitted.

A lot of language-fixing slacktivism easily becomes a trend because it's highly visible and requires no genuine effort. It's complete bullshit, though. This particular bullshit is awful since it trivializes all this historic pain to some inane game of word association. There's no point arguing, though, just drag your heels until they get bored and go off on some other crusade.

I'm always surprised when someone goes through more effort to oppose something than to accept the PR and build a little trust. We did this in a codebase of ours recently, it really wasn't a big deal to make the change and I don't understand why everyone throws up a big stink about it. If I'm going to spend my energy on something it seems like there's bigger things to tackle than pushing back against branch renames.

[dead]

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#255

Post by johnisgood » Mon, Feb 22, 2021, 3:17 AM UTC

Fogest wrote: post #239

johnisgood wrote: post #207

Earlier quoted context omitted.

These people (the genuine ones) lack context-awareness, so they could always find something racist, sexist, ableist, whateverist in anything you say. I just named something "blacklist" a couple of days ago. It has nothing to do with races, nor slavery. The only thing they have is a person in-between who connects the two because s/he is a troll, or because lacks context-awareness or whatever. I do not enable them.

Yeah I personally try and avoid the word police. If I am using a grammatically and contextually correct word I am not going to let someone tell me I can't use the word. What's hilarious to me is that I think the people who find there to be a problem with these words are the racist ones themselves. They are taking a word out of context and being offended by it out of context. They are the ones looking at a word and th…

> What's hilarious to me is that I think the people who find there to be a problem with these words are the racist ones themselves.

That is what I believe as well. I am not that preoccupied with races to make such associations. In all fairness, I still have no clue what "blacklist" has to do with races or slavery, even after having been "shown" this association, because it makes no sense given the context.

There might be people out there who genuinely do not know what the term means in the context of CS/IT, but then instead of assuming that everyone using the term is racist, they should try to figure out the actual meaning.

Killing a child or a parent in the context of computer science is something completely different than... actually killing a child or a parent. :) I can imagine "just kill the child" being misinterpreted, but that can only be the case if someone takes it out of context. Once the context is known, there should be no confusion or mystery.

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#256

Post by onethought » Mon, Feb 22, 2021, 3:19 AM UTC

sago wrote: post #131

ben509 wrote: post #50

Earlier quoted context omitted.

A lot of language-fixing slacktivism easily becomes a trend because it's highly visible and requires no genuine effort. It's complete bullshit, though. This particular bullshit is awful since it trivializes all this historic pain to some inane game of word association. There's no point arguing, though, just drag your heels until they get bored and go off on some other crusade.

I'm disagreeing with this thread. So I guess prepping for the down votes. > it trivializes all this historic pain to some inane game of word association I think addressing one issue does not 'trivialise' everything else. It can be horribly overstated but changing small fry can help provide momentum to change something more significant. > It's complete bullshit, It is obviously associating racism with something not in…

> But does it not strike to you at all problematic, at the moment, that 'black' is a synonym of 'bad' and 'white' of 'good'?

The inverse solution to this stop calling people with darker skin "Black"?

I kind of agree with what you're saying, but also it isn't true:

- Blacktie

- Black friday

- Blacktop

- Back in black

- Orange is the new black

- Men in black

All of those use black in a positive connotation. In my head black/white list has always been a metaphor for light/no light. As in black list can't see, white list can.

In asian Culture white is often associated with the colour of ghosts / the dead... So it'd get awkward if we keep modifying language to suit cultural references especially as English is a global language and while slavery was(is) a global problem it persists only really in the US as a current popular cultural issue.

In replying to you I've convinced myself at least the solution here can't possibly be to change all of the possible references of black to something else... it verges on new speak.

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#257

Post by pishpash » Mon, Feb 22, 2021, 3:21 AM UTC

dehrmann wrote: post #121

andrewzah wrote: post #5

“ The researchers now suggest that social-media platforms should incorporate chess language in their algorithms to avoid future incidents like this.” Or maybe don’t automate banning YouTube channels via nebulous black boxes? Or can advertisers stop being afraid of anything and everything?

Setting this case aside because it's sloppy reporting based on what another classifier found, this is a hard problem. Youtube gets too many videos and comments for humans to manually review them. Flagging helps, and Google likely uses ML more complex than just text and classifier (who is commenting, is the post trending, patterns in comments, etc.) When things do reach manual review, remember that humans make mistake…

Is it more complex than a text classifier though? If it truly took into account who is commenting, and context, etc., it would have noticed very few chess videos were being flagged for what seemed like racist language.

It seems more than likely somebody put in a very crude black box to process text without context. (Or should I say, opaque box, so the horde doesn't flag me?)

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#258

Post by mr_toad » Mon, Feb 22, 2021, 3:22 AM UTC

sago wrote: post #131

ben509 wrote: post #50

Earlier quoted context omitted.

A lot of language-fixing slacktivism easily becomes a trend because it's highly visible and requires no genuine effort. It's complete bullshit, though. This particular bullshit is awful since it trivializes all this historic pain to some inane game of word association. There's no point arguing, though, just drag your heels until they get bored and go off on some other crusade.

I'm disagreeing with this thread. So I guess prepping for the down votes. > it trivializes all this historic pain to some inane game of word association I think addressing one issue does not 'trivialise' everything else. It can be horribly overstated but changing small fry can help provide momentum to change something more significant. > It's complete bullshit, It is obviously associating racism with something not in…

> But does it not strike to you at all problematic, at the moment, that 'black' is a synonym of 'bad' and 'white' of 'good'?

Black is associated with the night and darkness and death in many cultures. Fear of the darkness is common to all humans.

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#259

Post by Sebb767 » Mon, Feb 22, 2021, 3:22 AM UTC

Retric wrote: post #208

dogma1138 wrote: post #189

Earlier quoted context omitted.

Over 500 hours of video are uploaded each minute so yeah... even 100,000 people might not be enough to just deal With flags...

> 100,000 people might not be enough to just deal With flags... You’re off by orders of magnitude. 500 * 60 * 24 * 365 = 262800000 hours per year / 1.3x playback speed / 2,000 hours of work per person per year = ~101,077 people to watch every uploaded video. Less if people are watching at higher playback speeds or working significantly more than a 40h week. I would personally be surprised if even 1 in 1,000 videos wh…

> 500 * 60 * 24 * 365 = 262800000 hours per year / 1.3x playback speed / 2,000 hours of work per person per year = ~101,077 people to watch every uploaded video.

When you have people watch the video at that fast speed and instantly make a decision, you might as well keep the algorithm - the failure rate won't be too different.

Additionally, these people need holidays, managers, infrastructure and an actual flag will probably not instantly be clear - you'll need to point out the specific offence and make a case. You might even need to look things up. Then you'll have to handle appeals and discussions, because changing the AI blackbox to an appeal-less human black box wouldn't improve the situation much[0]. Next, you need people who can understand the specific language of that video - most will be english, but what happens when a Nigerian video is flagged? So you need people from that country or at least familiar with it. Plus infrastructure.

Overall, you're probably looking at something approaching 400 or 500 people, with this very low flag rate. Assuming conservative 30k$/year [1], this is 15 million USD alone. Doable, yes, but its not the no-brainer you make it out to be.

[0] See this case with FB, which was a human error: https://www.nbcnews.com/tech/social-media/facebook-under-fir... [1] Which is probably optimistic for reviewers, but we'll have managerial infrastructure to pull that average up.

Re: YouTube blocked chess channel after mistaken detection of racial slurs

#260

Post by pishpash » Mon, Feb 22, 2021, 3:23 AM UTC

bartread wrote: post #62

andrewzah wrote: post #5

“ The researchers now suggest that social-media platforms should incorporate chess language in their algorithms to avoid future incidents like this.” Or maybe don’t automate banning YouTube channels via nebulous black boxes? Or can advertisers stop being afraid of anything and everything?

Bear in mind what we have here is an example of the cutting edge of AI in its real world use. It doesn't exactly inspire faith in AI if some of the best is this thick when it comes to making decisions that would be incredibly basic for a human of even below average intelligence.

Are you sure a human, never mind one of below average intelligence, gets this? We have people right now in Silicon Valley calling for renaming CS words because they sound racist to them.