AI Won't Save Us From Fake News: YouTube's Fact Checking Tool Thinks Notre Dame Fire Is About 9/11
from the content-moderation-doesn't-work-this-way dept
In the ongoing moral panic about social media algorithms and what they recommend, there are various suggestions on how the companies might "improve" what they do -- and many of them suggest relying on newer, better, algorithms. It's the "nerd harder" approach. Mark Zuckerberg himself, last year, repeatedly suggested that investing heavily in AI would be a big part of dealing with content moderation questions. This has always been a bit silly, but as if to demonstrate how silly this notion is, yesterday, during the tragic fire at Notre Dame Cathedral in Paris, YouTube's fancy new "fact checking AI" seemed to think multiple videos of the fire were actually referring to the September 11th, 2001 attacks on the US and linked to a page on Encycolpedia Britannica with more info about the attacks:
These links didn't last for very long, but at the very least, it should be a reminder that expecting AI to magically fact check breaking news in real-time is (at the very least) a long, long way off, and at worst, a nearly impossible request.
This puts YouTube and others in an impossible position of their own. Just a few weeks ago, people were freaking out that YouTube and Facebook (briefly) allowed videos from the attack in Christchurch to be on their platforms -- and have been demanding that the platforms "do something" in response. Having a tool that provides at least some sort of context, or even counterpoint to nonsense (when people start posting nonsense) certainly seems like a good idea. But it requires a level of sophistication and accuracy that is currently severely lacking.
One response to all of this would be to admit that human beings are not perfect, that social media sometimes reflects all aspects of humanity, and that sometimes bad stuff is going to make it online, but that doesn't seem acceptable to a large number of people. Given that, they're going to have to accept that sometimes AI is going to get this kind of stuff laughably wrong.
Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.
Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.
While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.
–The Techdirt Team
Filed Under: ai, fact checking, fake news, notre dame fire
Companies: youtube
Reader Comments
The First Word
“Re: Actually "Gary" we only want YOU to be decent and not steal.
Are you kidding me.
If a person gets mugged in Walmart, you go after the mugger, not Walmart.
If someone steals a TV from a car at Best Buy, you go after the theif, not Best Buy.
If someone is selling cocaine near the candy aisle at Walgreens, you arrest the dealer, not Walgreens.
Hell, if the pharmacist is selling painkillers illegally at Walgreens, YOU ARREST THE ACTUAL CRIMINAL.
Why is the concept of personal responsibility so antithetical to your beliefs, especially since you have long spouted your SovCit ideologies?
made the First Word by Ninja
Subscribe: RSS
View by: Time | Thread
This is reaching a bit. People know AI (or autocorrect, for that matter) will make errors. It's not a big deal.
Content moderation is necessary because of those who abuse content, or who allow piracy. "Good" people standing down while bad actors perform malicious acts are why these regulations were necessary. Section 230 is on the way out because platforms converted it from a shield into a sword.
The internet is not some magical place where laws cease to exist, just because they can easily be broken online. The first priority will be to stop the lawbreaking, and only then whatever is left over can be devoted to freedom of expression.
[ link to this | view in chronology ]
Re:
I'm not even going to bother. You're beyond parody.
[ link to this | view in chronology ]
Re:
We already have laws against all of the illegal activity you allude to. We really don't need more that say "oh, it's also still illegal when done on the internet" and we certainly don't need to deputize (and burden) every web site on Earth to police the public.
These regulations were not necessary and Section 230 is only being fought against because without it people could sue the platforms for the actions of the public since the platforms have all the money; It's not profitable to sue individuals, the actual law breakers.
[ link to this | view in chronology ]
Re:
"The first priority will be to stop the lawbreaking"
Then why do you constantly argue for the mechanisms that make this easier to be dismantled?
[ link to this | view in chronology ]
More like it is not acceptable to a very vocal minority, and the politicians who listen to them. Do nothing has not been acceptable to politicians for a long time, and that gives vocal minorities a big lever, as the alternative to their desire is do nothing, and tell them they need not follow links, and can click away from anything that they do not like.
[ link to this | view in chronology ]
Re:
"Do nothing has not been acceptable to politicians for a long time"
lol - sure
Ever heard of the Do Nothing Congress?
When was the last time this nation had a real budget?
[ link to this | view in chronology ]
Re: Re:
And yet the 'do nothing' Congress still manages to pass a butt ton of mostly bad laws each year, usually meant to appease a vocal yet influential minority.
[ link to this | view in chronology ]
That's fine. You have 1 hour to fix your or your AI's mistake. That should be plenty.
/s (just in case)
[ link to this | view in chronology ]
Re:
Twenty years of intransigence by internet platforms has caused this.
[ link to this | view in chronology ]
Re: Re:
Wrong again. Cranio-rectally inverted politicians and their adoring sheep caused this. The platforms have done nothing wrong.
[ link to this | view in chronology ]
Re: Re: Re:
Keep telling yourself that. The #metoo predators used to think the same thing.
[ link to this | view in chronology ]
Re: Re: Re: Re:
Critical thinking tells me that. An ability those of your ilk seem to lack.
[ link to this | view in chronology ]
Re: Re: Re: Re:
Since you’re the king of projection Jhon boy. How many people did you sexually assault?
[ link to this | view in chronology ]
Re: Re:
Twenty years of intransigence by internet platforms has caused this.
You mean 20 years of our ISP's not monitoring our every activity?
You are demanding - in no uncertain terms - to have your every private email opened and read just in case it contains forbidden material. And encrypted material blocked by default, just to be safe.
Every post you make, every email, every picture you take with your cellphone.
Some will shout, "Aha! I don't have a cellphone." But you are still posting here, and DEMANDING that you be censored.
Good job AC.
[ link to this | view in chronology ]
Re: Re: Re:
E-mail is private. These measures deal with what is posted publicly.
The third-grade logic is amusing.
[ link to this | view in chronology ]
Re: Re: Re: Re:
If you really believe that then you're far more naive than you know.
[ link to this | view in chronology ]
Re: Re: Re: Re:
Except when illegally spam them as your business model bro.
[ link to this | view in chronology ]
Re: Re: Re: Re:
E-mail is private. These measures deal with what is posted publicly.
I have seen many comments that calmly explain that everything should be checked, because it can be checked. Because - pirates bad.
Are you really so naive that you don't think ISP level filtering wouldn't check email too? Or just good with it because only the pirates will be censored, eh?
[ link to this | view in chronology ]
Re: Re: Re: Re:
"E-mail is private."
Lol.
I bet you think you're perfectly anonymous here, too.
[ link to this | view in chronology ]
Re: Re:
No, the 20 years of the internet bending backwards spending the GDP of small nation on the demanded filters, yet being attacked because they haven't performed literal magic has caused this., Well, that and the ignorance of those like you, who still don't understand what reality looks lie.
[ link to this | view in chronology ]
Re:
You have 60 minutes to comply...
[ link to this | view in chronology ]
Actually "Gary" we only want YOU to be decent and not steal.
No. We just want you acting same as in real life, WITHOUT Section 230 essentially legalizing piracy and vile attacks instead of reasonable discourse.
Why does The Internet require an exception to civil society?
Do you think that shouting over every little irritation is progress?
(Of course, it's a given that you DO think piracy is okay up to justified, and that links to infringed content is "free speech" that must not be suppressed. Skip that.)
Now, what's the purpose of Section 230? -- To make it easy for The Public to Publish without corporate or gov't interference.
But Rule by corporations over 1A Rights is what YOU advocate with Masnick's view of Section 230. -- And NO, "moderation" is not same as the absolute arbitrary control that masnicks want so can simply suppress all opposition to corporatism. -- And there is NO "separate but equal" comparing giant platforms to tiny ones that Google doesn't have to index, so will never be discovered.
Section 230 makes individual Publishers. The corporations are NOT the Publishers, not liable, but they intend to keep Editorial Control too. -- That's the real de facto censorship, kid, NOT the coming attempts to limit piracy and have a snippet tax.
In other words, you, Techdirt, and especially Masnick have it all backwards.
-s-u-b-s-t-i-t-u-t-e -h-o-r-i-z-o-n-t-a-l -r-u-l-e
Oh, and especially for you, "Gary", who's actually Geigner: we want you to quit astro-turfing.
[ link to this | view in chronology ]
Re: Actually "Gary" we only want YOU to be decent and not steal.
Are you kidding me.
If a person gets mugged in Walmart, you go after the mugger, not Walmart.
If someone steals a TV from a car at Best Buy, you go after the theif, not Best Buy.
If someone is selling cocaine near the candy aisle at Walgreens, you arrest the dealer, not Walgreens.
Hell, if the pharmacist is selling painkillers illegally at Walgreens, YOU ARREST THE ACTUAL CRIMINAL.
Why is the concept of personal responsibility so antithetical to your beliefs, especially since you have long spouted your SovCit ideologies?
[ link to this | view in chronology ]
Re: Re: Actually "Gary"
Why is the concept of personal responsibility so antithetical to your beliefs, especially since you have long spouted your SovCit ideologies?
The core belief of the Sod-Cit is they aren't responsible for anything. Submitting to others laws is weakness. Therefore Blue-Balls hates getting downvoted because we show power over him.
[ link to this | view in chronology ]
Re: Re: Re: Actually "Gary"
It is worse than that - they want laws to not apply to them but to apply to other people - and their concepts of the law are by definition utterly delusional.
[ link to this | view in chronology ]
Re: Actually “we” want you to keep your promise and leave
Who’s this we shit bro? You ain’t French.
[ link to this | view in chronology ]
Re: Actually "Gary" we only want World Peace
Actually Blue Balls copying is not Theft. You'd know that if you were paying attention.
Blocking suspicious content is censoring. Or is it only censorship if you get blocked?
You are still demanding global, automatic censorship on a mass scale. Nothing less. Every Communication to be checked in case it might be owned by a corporation.
Corporate censorship backed by common law!
[ link to this | view in chronology ]
Re: Actually "Gary" we only want YOU to be decent and not steal.
"We just want you acting same as in real life"
The way it works in real life is that if someone misuses the phone, you go after the caller, not AT&T.
Why do you insist the internet not be held to the same logic.
[ link to this | view in chronology ]
AI will continue to result in embarrassing failures, such as Google's labeling of a notorious Nazi concentration camp as a "jungle gym" and identifying people of black African descent as "gorillas."
And in what could be the key to future of AI "improvements" -- Google's solution to accidental algorithmic racism: ban gorillas.
https://www.theguardian.com/technology/2018/jan/12/google-racism-ban-gorilla-black-people
[ link to this | view in chronology ]
Hosting filters
Well, this certainly puts my mind to rest as to how well the EU's Copyright Directive hosting filters are gonna work.
Oh yes indeed
[ link to this | view in chronology ]
Re: Hosting filters
No worries, they just need to Nerd Harder, and so long as they're really trying then it will be trivial to solve. I mean really, if a politician can (tell someone else to) do it then an actual expert should have no problem.
[ link to this | view in chronology ]
The trouble with AI
The trouble with AIs is that a sufficiently evolved AI is indistinguishable from a human. Meaning it will have its own biases and believes. It won't be objective, at all.
Which again makes it completely useless for exactly that which most of the people think it could be used.
[ link to this | view in chronology ]
I believe there may be one point where we manage to create AI smart enough that it will be able to read into context, nuances and the likes and do filtering right. None of us reading this will probably be here to see it happen.
Even then, we know AI will be as biased as their learning data sets and maintainers are. We'll probably develop fully autonomous and sentient androids before we can do filtering right.
[ link to this | view in chronology ]
Re:
If people can't figure out how to read context and nuances correctly, there's absolutely no way an AI will be able to do it.
[ link to this | view in chronology ]
The world was a much better place when we expected everyone to be thick skinnned, and we didn't police words and thoughts to appease sensitive liberals.
[ link to this | view in chronology ]
Re:
Too bad the days of suck it up, butter cup, here's a tissue for your issue are long behind us now.
[ link to this | view in chronology ]
Re: Re:
It's like they think it's free. I bet our current coddling of emotional people still not over being wrong about an election has cost our economy an easy trillion dollars.
[ link to this | view in chronology ]
Re: Re: Re:
She lost. Get over it bro.
[ link to this | view in chronology ]
Re:
Like last week when you ran away because someone hurt your feel-feels?
[ link to this | view in chronology ]
Are we so sure the AI got it so wrong?
What is the primary defining feature of the 9/11 attack? They were Islamic in origin.
What are the primary features of most church desecrations and attacks in Europe? They're Islamic in origin.
What is the primary feature of the Notre Dame Fire? Well, we're not yet sure that it was of Islamist origin, but given the mountains upon mountains upon mountains of previous evidence of Islam destroying any monuments that don't match their culture/religion, it's pretty reasonable to guess that this might be another case of Jihadis Gone Wild.
Just look at the Bamiyan Buddhas. Not Islamic, so destroyed by Islamists. Priceless ancient monuments and knowledge of history, gone forever.
[ link to this | view in chronology ]
GIGO
[ link to this | view in chronology ]