By removing “extremist content,” platforms are purging human rights evidence
nytimes.com
nytimes.com
Archival of footage shouldn't be dependent on a third party platform that can, at will, delete every single video on a whim. Likely for this sort of thing a diligent archivist will need to put in the effort of storing it safely, because those videos will almost all end up being deleted on platforms that rely on ad revenue.
There are 110,000 public libraries in the United States. Plus thousands more private libraries. Seems like enough for U.S. history, at least.
Maybe that's a problem that everyone has? But it would be best if there was some offshore, FBI proof, kind of way to do this archiving.
I wouldn't be so sure about that. After 9/11, the libraries were the first to stand up to the government's desperate search for information. The feds wanted everyone's lending history, and ended up backing off when the libraries put up a fight.
At some point, you need curators and archivists.
To my understanding websites can chose to not be included in the Wayback Machine.
----> Why isn't the site I'm looking for in the archive?
Some sites may not be included because the automated crawlers were unaware of their existence at the time of the crawl. It's also possible that some sites were not archived because they were password protected, blocked by robots.txt, or otherwise inaccessible to our automated systems. Site owners might have also requested that their sites be excluded from the Wayback Machine.
----> How can I exclude or remove my site's pages from the Wayback Machine?
You can send an email request for us to review to info@archive.org with the URL (web address) in the text of your message.
https://help.archive.org/hc/en-us/articles/360004651732-Usin...
I feel like we’re in the Internet equivalent of the early years of Hollywood. We’ve lost so many precious early movies because there were no backups of the film.
[0] https://blog.archive.org/2017/04/17/robots-txt-meant-for-sea...
But to find a provider of public access to archives more expansive than YouTube's, well, that's improvement.
I somehow don't get the same impression from same impression from YouTube
I'm sure if they started doing this the Twitter outrage warriors wouldn't be happy about it and add to the difficulty of their job.
Somebody should classify this as an addictive behavior.
'The immanent urge to get emotionally excited by interacting with an equally excitable peer group on issues only remotely related to one's personal life'.
Internet Archive seems to be handling this kind of trouble quite well.
Goodreads will list any book for the time being, even ones banned from Amazon. One example is The Culture of Critique - the link to Amazon goes to a broken page, but links to buy it from other vendors remain.
Link to the book - https://www.goodreads.com/book/show/182136.The_Culture_of_Cr...
Liveleak, seriously? I'm not convinced that a site known for hosting gory and violent videos is really what these human rights activists are looking for.
[Edit] Some context: https://www.businessinsider.com/profile-of-hayden-hewitt-fou...
People who want to watch violent videos != human rights activists. There's the "blood and gore" people, and the "social justice warrior" people, and not a ton of overlap.
In order to solve this use case there would really need to be a site that combines LiveLeak's content moderation policy with an activist ethos.
So they're not really a great example.
You mean gory and violent videos like war crimes and human rights violations? Sounds like exactly the right site.
Thankfully censorship-resistant anonymous p2p file sharing systems exist, I hope they become more popular for archival work.
"You should not show identifiable people being beaten!"
That's true, because it could go one way or another depending on whether the revolution succeeds. If it does, though, it can help identify the perpetrators of brutality.
Publishing it is irresponsible, but destroying them is as well. This is potential criminal evidences.
Across prestigious universities, mini-think tanks that study "tech and society" have been popping up. As far as I can tell, they exist solely to contribute to this tidal wave of narrative. They look at what is happening, and make up a reason it's bad. I can't think of a single example where they concluded that tech did something good.
I suspect they are partly worried about results being gamed there...but everything is gamed. Maybe the main search results should be empty too.
Google suggestions for "is stephen miller":
* is stephen miller married?
Bing suggestions for "is stephen miller":
* is stephen miller gay?
* is steven miles married?
Maybe there's location stuff going on, or maybe they're guessing at what you're searching for based on previous searches or something.
YouTube is better off with meta tag recommendations instead of their fancy little AI.
I sort of feel like their recommendation AI is a bunch of if-then statements designed to funnel you into more heavily-monetized content if they think you display any interest at all.
I'm a guy in my 30s. Nothing in my view history ever says, "Hey, this guy is progressive enough to wear makeup." At that, I watch english, polish, and french videos. Ads I get, Spanish. I dont know spanish. I'd accept general language learning ads. But tostitos and I think one was health or house insurance... in spanish. It's like the ad bot is some redneck that thinks there are two languages, American and not-American.
The recommendation bot and ad bot are just terrible. The moment theres a half way decent competitor to youtube, my ass is jumping ship.
Reality TV recommendations... Jersey shore... celebrity gossip...everywhere...it took 2 weeks to disappear. Strep throat is cured in less time than contracting celeb gossip videos.
Your recommendations may return to normal.
Although one person did complain Google closed his account because he used that app: https://news.ycombinator.com/item?id=21247759
I do think that they are starting to lose the power-user userbase when it comes to browsers and search engines. But their email, docs/sheets/whatever, drive, and now their DNS registrar services are all cemented into business flow.
It makes sense. They don't need the power users - they need the most users. As long as they have the most users, businesses (where the power-users are working) are going to use ancillary services like Google MyBusiness, Google Analytics, and embrace google-backed (and ranking-factor) tech like SSL certificates and progressive web apps.
I'll bet it's dwarfed by the number of times I get useful results from Google and total garbage from DDG.
Earlier today I searched for "CADT syndrome" to figure out what another HN comment meant by it. Both engines automatically correct that to "cast syndrome" if you search without quotes. But only Google provided a relevant result after being corrected "no, I really meant CADT syndrome". DDG gamely provided the "cast syndrome" results after correcting the query.
I primarily use DDG, and in my experience when its results are bad, Google's usually are for that query too.
Two, the others are referencing controversial opinion content in particular. They are saying that if the search is about something controversial or politically charged, the google result is noticeably lacking. This seems more concerning to many of us because this isn't just a newer search engine not quite up to par, this is the search engine purposefully pushing users into filter bubbles of Googles choosing!
I hope I cleared up the difference between what you are referring to and what the others are talking about. Anecdotally, I can confirm as a heavy youtube user (both web and youtube-dl/mps) that youtube search is also affected. Many searches for things that aren't even that "out there" return no results, even with the exact name of the video. If you don't have the actual link from somewhere else, you can't find it. I think what google is doing is dangerous and a very slippery slope. Many of my datahording friends have been on youtube downloading sprees because they are afraid google is about to memory hole important videos that don't exist anywhere else that is easily accessible.
ps. This is also the nice part of the bangs of ddg. Don't like the ddg results, just do it again with !gyear etc. (vs the google gui way, and I don't even know if googles daterange: operator even works anymore)
As of now, the best resource (https://www.jwz.org/doc/cadt.html) is avaiable on both search engines as the first result for this query, and somewhere in the first ten or so results for "CADT Syndrome". The best set of results (for my Google history) is in Google's verbatim mode searching for "CADT syndrome" (the link under the search bar with the autocorrected version) which includes mailing lists and release notes from various opensource projects
interestingly there is also another meaning of CADT (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4610895/) which is less favored in google, but shows second place in ddg. Perhaps if we were into medicine papers this would show up higher in google.
Only ddg results include this (Hacker News) page and the comment you referenced (https://news.ycombinator.com/item?id=21337404)
Bottom line, both gave me useful results, ddg needs a better verbatim mode, google tracks me
What is that alternate platform?
For many people, web video is YouTube.
But YT is where channels I watch such as Linus TechTips, BigClive, Cody's Channel, Chef John, Glen and Friends, Andrew Camarata et al publish their content to.
Realistically, how do we change this?
Millions upon millions of people watching YT, but GP "needs to get out more" in order to experience your fantasy platform that doesn't exist. Ok.
Ask 100 random people on any city street where they'd find videos on the web and I am betting 0 of them will come up with any answer resembling yours.
> A not-yet-existent, co-op-owned platform where
Yes, so "not-yet-existent" answers the question. The alternative does not exist.
But you seem like you have an idea fleshed out, so why don't you build this alternate video platform?
This is how every platform starts out. Success is a curse that drags companies into politics and money games, causing them to succumb to censorship and other unintended policies.
The only solution is a platform that can prove that it has no ability to censor content. This requires advanced cryptography knowledge. Never trust the "word" of a company or a platform.
If I set up a machine under my desk at home with a domain (courtesy of dynamic DNS), and provide a few terabytes of free-of-charge storage to a small bunch of people, am I "erasing history" if I can't keep all their stuff forever?
Or do I have to be a Youtube-sized corporate entity before I'm a fucking asshole?
If you say "hey guys, feel free to use this storage to host stuff" and then delete it, then yes, you're an asshole.
If you say "hey guys, I have some hard drive space, send me stuff you think I might like and if I like it I'll host it for a while before deleting it" then no, you're not an asshole.
If you say something that sounds like the first statement but is actually the second statement because of some obfuscated legalese attached to a footnote, and then delete it, then you're back to being an asshole.
If you want things archived, non-profit solutions exist, or you should be doing it yourself and making sure the content is shared among people who are also able to archive it.
Because they claimed they cared about this sort of thing. For a while, it seemed as if they meant it.
You seem to be under the impression that non-profits don't ever go out of business, which is completely untrue.
I'm not saying that for-profits are the solution, but maybe something like a trusted .gov or network of trusted .edus would be better.
This is modern day cynicism. I fully understand and am unsurprised by their lack of good storage, but we shouldn't help companies lower the bar of their own existence to be purely profit driven entities.
But it's not all a corporation is.
internet media in the current form is subject to great purges of intentional and accidental nature. no matter the corporate intention, there is a great danger of major events being deleted from social memory, and the engineering attempts currently in play, will easily become best practice in gov parlance.
...Just a bit undermined by the screenshot of Ogrish video titles in the article... https://media.businessinsider.com/images/54117ca9ecad04406f2...
Honestly, why do platforms need to delete anything. Why not just have options like most search engines sort have... Let the users decide if they want to see objectionable or even violent content. They are adults they can make up their own mind.
Consider duckduckgo's search results you can choose:
Strict
*No Adult content*
Moderate
*No Explicit images or video*
Off
*Don't filter content*Future is bleak if we do nothing.
Or something better.
Keep refining and optimizing the system.
Tech Companies Are Deleting Evidence of War Crimes (theatlantic.com) https://news.ycombinator.com/item?id=19864994
This presents an opportunity for another Internet Archive-like repo as mentioned in earlier comments. Unfortunately the barrier to entry (just getting people to know about it) would be super high.
The WITNESS movement associated with one of the authors is worth checking out: https://www.witness.org/
Just throwing human moderators at the problem probably has more cost than just the salaries involved.
The alternatives are
1. Algorithms, which always going to fail in human rights type "edge cases" like the ones described in the article, and which really shouldn't be seen as edge cases IMO, or
2. Bury your head in the sand and allow terrible content to exist.
Do you have any better ideas for how to deal with processing traumatic content at scale?
If I understand your position: you don't have any better ideas, and you're okay with using ML even if it means deleting evidence of human rights violations. Is that accurate?
I'm not particularly sympathetic to the notion that YouTube should be obligated, either legally or ethically, to host content in perpetuity. OTOH, if rapid filtering results in such content being taken down so quickly that even archivists and reporters don't have a reasonable chance at identifying and saving it, then I do see that as a potential ethical (but not legal) dilemma, and an interesting technical challenge.
My initial thought upon reading the title was that this story--such as it is--was more a commentary on the unintended consequences of the censorship of terrorist-related propaganda videos. From that perspective, I don't think the distinction between ML and human filters is that important; the more interesting issue seems to be the scope and creep of censorship and how we can expect tech companies to draw equitable lines within modern political cultures that suffer from hyper-partisan political discourse. It's not hard to imagine a situation where simply permitting any atrocity videos will be seen as partisan--supporting one side or another by promoting particular narratives.
There is also clearly an ability to follow direction or enforce appropriate (e.g. the janitors of /b/ /pol/ /g/ and /fit/ are clearly all operating on different sets of rules).
I am 100% in agreement that you cannot hire an army of temps off the street and tell them to moderate the internet, but it's clearly not something beyond human capability. You just have to be hiring in the right places.
i.e. they will typically use contractors or contract houses for moderation work. Why? Because it allows for opportunities to minimize the cost of counseling (If you're doing serious moderation, they must provide mental health support services. But fixed-term contracts make this easier to contain/control, at the cost of the employee.)
Maybe they send the data to law enforcement automatically? (I've heard that they do that for child pornography; maybe they do, or could do, the same for terrorist material as well.)
It's naive to think these videos are deleted. Homeland Security shows up and YouTube has no problems giving them the videos. (They probably have an entirely separate search interface that does nothing but surface that extremist content for the Feds. So Homeland Security probably doesn't even have to leave their office.)
Doesn’t stop countries from trying. The CLOUD Act in the US is the US trying to extend its influence into other countries.
YouTube is trying to straddle the line between platform and publisher and it feels like the hammer will soon drop.
It's not a volume problem though, because the videos are never removed. They are just effectively delisted. Homeland Security shows up and I guarantee YouTube hands the videos over to them. So YouTube has plenty of space.
Of course that also means that the original complaint of YouTube destroying evidence is also bogus, but answering a bogus complaint with a bogus excuse as the pro YouTube crowd seems to be doing doesn't really get us anywhere.
Video from Tim Pool about it: https://youtu.be/tzTod38VS8c
At [1] the quote is:
Hillary Clinton waded into the Democratic primary on Friday by suggesting that Russia was “grooming” Representative Tulsi Gabbard of Hawaii as a third-party candidate for president.
By the next day at [2] it had changed to:
Hillary Clinton waded into the Democratic primary on Friday by suggesting that Russia was backing Representative Tulsi Gabbard of Hawaii for president and that Republicans were “grooming” her as a third-party candidate.
The NY Times published a correction:
Correction: Oct. 23, 2019 An earlier version of this article described incorrectly an element of Hillary Clinton's recent comments about Representative Tulsi Gabbard. While Mrs. Clinton said that a Democratic presidential candidate was "the favorite of the Russians," and an aide later confirmed the reference was to Ms. Gabbard, Mrs. Clinton's remark about the “grooming” of a third-party candidate in the 2020 race was in response to a question about the Republicans’ strategy, not about Russian intervention.
This matches other reporting of the same interview. Politco reported it like this[3]:
Clinton, the 2016 Democratic nominee, made the comments on the podcast hosted by David Plouffe, who managed former President Barack Obama’s 2008 campaign. Clinton also said that Republicans could boost Gabbard to become a third-party “spoiler”
Conclusion: The NY Times did the correct thing in updating their reporting, and in publishing the correction. The reporting is now more accurate that it was initially.
[3] https://www.politico.com/news/2019/10/18/hillary-clinton-tul...
No, the correct thing would have been to not publish fake news in the first place. Oh, but they said whoops, so I guess all's well that ends well?
But yeah, people do like easy answers rather than digging into complexity.
They're erasing a good chunk of the present as well. It is widely known and observable that they algorithmically throttle and aggressively demonetize anything even remotely conservative in the US. E.g. Tim Pool (a liberal who nevertheless likes to point out the lunacy and hypocrisy of the leftist "free press") often gets demonetized under the flimsiest of pretenses.