I do want to mention one more thing. There are more leaks that will be shared in the coming days. Many of which are related to mass censorship, theft, fraud going on at Meta.
This is just one leak related to censorship and Israel. There are still more.
ICW was created with the purpose of exposing corruption related to government and anything tech related. We focus on high quality investigative work. By quality, I mean actually running experiments and going beyond surface level analysis. While everyone has their own biases/leanings, we try to separate our investigations from our opinions (which we do provide in a clearly labeled separate section, usually called "Discussions").
As you can see from our post timelines, I spend months doing this kind of investigative work. And it takes 100+ hours to do a proper investigation, let alone write the actual report.
You can check out our previous investigations here:
The Youtube Algorithm and Manufacturing Consent (https://archive.org/details/youtube-icw)
- We collected the worlds largest Youtube recommendation dataset using a custom built watch bot. We concluded that almost all Youtube users are 1-click away from far-right radicalizing videos.
What can we learn from the Andrew Tate data breach (https://archive.org/details/tate_data_breach/)
- We looked at the leaks of Andrew Tate's school and calculated through simulations exactly how much money he was making from the project, and allude to what this may mean for his taxes. We also run a state-of-the-art analysis into what kind of posts people make there, as well as survey the user demographics.
We do not currently have a website. And have had $0 funding so far. So entirely out of pocket. From the beginning we opened donations via BTC and ETH, but didn't receive anything yet.
I am NRU (alias) the lead investigator with a background in AI. I am currently driving the entire investigation for these projects. I occasionally collaborate with Drop Site News and BBC for some of the work we do.
There are several indications that this is a single person “organization”, and not a true research organization. And that’s okay if the research is actually factual and grounded, but the misrepresentation can be a bit off-putting to some, and many of the methodologies seem flawed for the conclusions drawn. There is a distinct lack of actual scientific rigor that seems to be brushed over with language like “state of the art” and “world’s largest” without any evidence those statements are actually true.
I would recommend considering these findings to be moderate to highly unreliable and entirely unverified due to the lack of scientific rigor. Effectively, it is best to consider this as potential political propaganda, which doesn’t indicate any sort of nefariousness, only that it may be motivated by an agenda to prove something rather than independent analysis.
Massive respect to dang et al, but a blind man can see that posts about Gaza get brigaded.
We would all love to see a report on unusual activity in regards to posts on Gaza.
I m not surprised a bit by this. Meta has again and again shown that they have no scruples and no moral compass. But it's surprising how little knowledge about it and reaction there is from its users, who are otherwise very very critical when "mainstream media" does propaganda on them. It's probably because FB takes a very low key approach with branding and tries to make itself transparent to its users.
> Please don't post insinuations about astroturfing, shilling, brigading, foreign agents, and the like. It degrades discussion and is usually mistaken. If you're worried about abuse, email hn@ycombinator.com and we'll look at the data.
@dang and team
I think the community would be interested to know the activity around this post, including moderation efforts. I’ve been doing cursory refreshes and seeing what I would consider “brigading”, but I could just be paranoid.
Anyone else have the same questions? I’ll be emailing later and encourage others to do the same.
I emailed the moderator and asked them why this post was constantly getting flagged and locked.
Here is their reply (quote):
"Looks like user vouches outweighed user flags, at least for now."
I am not familiar with how hackernews works, but this is what ChatGPT returned on this:
"""
On Hacker News, not all users have the ability to vouch for others. Typically, vouching is restricted to users who have achieved a certain level of trust or reputation within the community. This means that only established users, often those with a higher karma score or a longer history of positive contributions, can vouch for others.
If you find that you cannot vouch for someone, it may be because you haven't met the necessary criteria set by Hacker News. The platform aims to maintain the quality of endorsements, ensuring that only credible users can influence the reputation of others. If you're looking to vouch in the future, focusing on contributing positively to discussions and building your karma can help you reach that level.
"""
Yes. The longest comment on this thread starts out by saying there's so many typos.
These might not be state actors, but like you said, a blind man can see that when someone's comment on genocide is "the article has so many typos", that person already started out with the conclusion.
"Israel ranks 3rd in the most posts targeted by TDRs out of any country.
"On a per-capita basis, Israel ranks 1st in the most posts targeted by TDRs, and has 3 times more TDR targeted posts per-capita than the country with the 2nd most submissions.
"Its important to contextualize this with the fact that almost all governments reporting to Meta primarily censor citizens of their own countries. Israel is the exception as only 1.3% of its takedown requests are actually targeted towards Israeli’s (14th most targeted country).
"For reference, 63% of Malaysia TDRs target Malaysian content, and 95% of Brazil’s TDRs target Brazilian content"
(Back to me): They are saying that Israel is doing this at a huge scale and, very unusually, almost exclusively targeting people in other countries.
The leakers speculate that Israel is doing this at at scale that is poisoning Meta's ML filtering inputs, so that they are now, without direct Israeli involvement, Meta is carrying on their own censorship of other countries. Since Israel's active TDR's produced the vast bulk of recent terrorism related TDRs, the leakers think that almost all of the 38 million Facebook posts that were censored automatically by Facebook ML for terrorism reasons during the period since the Simchat Torah attacks were because of this poisoning, but here they are drawing more on inference than direct evidence.
I had not realized the scale of the program here, or that they were targeting external countries like this, so I (someone reasonably familiar with Facebook's moderation) learned something from reading this.
Exactly, this study was on antisemitism but you can imagine other countries doing this for other content.
By overloading the Meta ML loop, they’re guaranteeing that the censorship is from meta and not them.
Antisemitism, political posts, rallies, anti-government sentiment, it’s an assault on free speech. Add the fact that news organizations have basically bent the knee to this US Administration and it’s pretty much game over for free thinking journalism w/ access. (Good journalism doesn’t necessarily require red tape access)
Many many countries are doing this, which is shown in figure 5. However, no other country comes close to the amount of censorship done by Israel. Additionally, most countries primarily do internal censorship. Again, Israel is the only country that is censoring other countries to this great extent. And almost all of it is related to the ongoing events.
Are we really shocked by this? It's screaming obvious but we still have many people whose head is buried deep in the sand.
Israel have claimed the title to start worlds first live stream genocide. It took them long time to get to where they are and know they will be forgiven.
It's crazy to think but is true, they have normalized killing of Palestinians.
Readit News
This is just one leak related to censorship and Israel. There are still more.
As you can see from our post timelines, I spend months doing this kind of investigative work. And it takes 100+ hours to do a proper investigation, let alone write the actual report.
You can check out our previous investigations here:
The Youtube Algorithm and Manufacturing Consent (https://archive.org/details/youtube-icw) - We collected the worlds largest Youtube recommendation dataset using a custom built watch bot. We concluded that almost all Youtube users are 1-click away from far-right radicalizing videos.
What can we learn from the Andrew Tate data breach (https://archive.org/details/tate_data_breach/) - We looked at the leaks of Andrew Tate's school and calculated through simulations exactly how much money he was making from the project, and allude to what this may mean for his taxes. We also run a state-of-the-art analysis into what kind of posts people make there, as well as survey the user demographics.
We do not currently have a website. And have had $0 funding so far. So entirely out of pocket. From the beginning we opened donations via BTC and ETH, but didn't receive anything yet.
I am NRU (alias) the lead investigator with a background in AI. I am currently driving the entire investigation for these projects. I occasionally collaborate with Drop Site News and BBC for some of the work we do.
You can contact us with:
- https://bsky.app/profile/icw-nru.bsky.social
- icw_nru@protonmail.com
I would recommend considering these findings to be moderate to highly unreliable and entirely unverified due to the lack of scientific rigor. Effectively, it is best to consider this as potential political propaganda, which doesn’t indicate any sort of nefariousness, only that it may be motivated by an agenda to prove something rather than independent analysis.
Dead Comment
@dang and team
I think the community would be interested to know the activity around this post, including moderation efforts. I’ve been doing cursory refreshes and seeing what I would consider “brigading”, but I could just be paranoid.
Anyone else have the same questions? I’ll be emailing later and encourage others to do the same.
Here is their reply (quote): "Looks like user vouches outweighed user flags, at least for now."
I am not familiar with how hackernews works, but this is what ChatGPT returned on this:
""" On Hacker News, not all users have the ability to vouch for others. Typically, vouching is restricted to users who have achieved a certain level of trust or reputation within the community. This means that only established users, often those with a higher karma score or a longer history of positive contributions, can vouch for others.
If you find that you cannot vouch for someone, it may be because you haven't met the necessary criteria set by Hacker News. The platform aims to maintain the quality of endorsements, ensuring that only credible users can influence the reputation of others. If you're looking to vouch in the future, focusing on contributing positively to discussions and building your karma can help you reach that level. """
Hopefully more users see your quote.
We would all love to see a report on unusual activity in regards to posts on Gaza.
These might not be state actors, but like you said, a blind man can see that when someone's comment on genocide is "the article has so many typos", that person already started out with the conclusion.
I have no proof, only suspicion. You can easily skirt “The Algorithm” using VPNs and get vastly different content.
"Israel ranks 3rd in the most posts targeted by TDRs out of any country.
"On a per-capita basis, Israel ranks 1st in the most posts targeted by TDRs, and has 3 times more TDR targeted posts per-capita than the country with the 2nd most submissions.
"Its important to contextualize this with the fact that almost all governments reporting to Meta primarily censor citizens of their own countries. Israel is the exception as only 1.3% of its takedown requests are actually targeted towards Israeli’s (14th most targeted country).
"For reference, 63% of Malaysia TDRs target Malaysian content, and 95% of Brazil’s TDRs target Brazilian content"
(Back to me): They are saying that Israel is doing this at a huge scale and, very unusually, almost exclusively targeting people in other countries.
The leakers speculate that Israel is doing this at at scale that is poisoning Meta's ML filtering inputs, so that they are now, without direct Israeli involvement, Meta is carrying on their own censorship of other countries. Since Israel's active TDR's produced the vast bulk of recent terrorism related TDRs, the leakers think that almost all of the 38 million Facebook posts that were censored automatically by Facebook ML for terrorism reasons during the period since the Simchat Torah attacks were because of this poisoning, but here they are drawing more on inference than direct evidence.
I had not realized the scale of the program here, or that they were targeting external countries like this, so I (someone reasonably familiar with Facebook's moderation) learned something from reading this.
By overloading the Meta ML loop, they’re guaranteeing that the censorship is from meta and not them.
Antisemitism, political posts, rallies, anti-government sentiment, it’s an assault on free speech. Add the fact that news organizations have basically bent the knee to this US Administration and it’s pretty much game over for free thinking journalism w/ access. (Good journalism doesn’t necessarily require red tape access)
Dead Comment
Dead Comment
Deleted Comment
Israel have claimed the title to start worlds first live stream genocide. It took them long time to get to where they are and know they will be forgiven.
It's crazy to think but is true, they have normalized killing of Palestinians.
This post is ranked very low, most likely it's getting lots of down vote.
Deleted Comment