Big Tech’s censorship is a menace to the free world. Who, and through what process, makes verdicts on posts? Ex-content moderator who was tasked by Facebook to “delete numerous conservative posts” answered these questions for our magazine.

 

――As a whistleblower, you have raised an alarm on Facebook’s censorship. First of all, would you be able to explain what you did at Cognizant, an IT company which was a Facebook subcontractor?

Hartwig: Yes. I started work there in March of 2018, and I was there for just under two years. I was a content moderator for Facebook and reviewed pieces of content one by one. For example, if someone was attacking someone else or calling someone a bad name, I would either delete it or leave it on the platform. We had to follow the policy guidance to decide if we should delete the material or leave it on the platform.

――I heard that there were about 1,500 content moderators at Cognizant in Phoenix.

Hartwig: That’s correct. It was about roughly 1,500.

――So, basically you were working as one of 1,500 content moderators at Cognizant and were allowed to decide which post should be censored or deleted based on the Facebook policy.

Hartwig: Yes, that’s correct. So we had to study the policy. The policy was very lengthy, very wordy, and sometimes I felt like a paralegal or a law professor because I had to study the policy and read it. It was a very difficult policy to understand, very nuanced, but we studied the policy to make decisions about whether to delete the content or leave it up. There’s a lot of gray area. There is some sense of subjectivity involved, and so sometimes content moderators have to make decisions on their own. If there was gray area or if we were uncertain, we would ask Facebook for guidance, so Facebook would give us instructions as to what to do. So we were always asking them for advice and trying to make them happy.

 

Facebook’s Double Standard

――And this past July, you went to Project Veritas and became a whistleblower.

Hartwig: Exactly. So in the summer of 2019, I reached out to Project Veritas because I noticed a few examples of bias against conservatives, against political viewpoints. Project Veritas, they do undercover journalism, and they help expose bias and censorship. They gave me a camera, and I started filming with a hidden camera. So I filmed for a total of about nine months. In the last couple of months, I was filming almost continuously, almost every day.

――In your Project Veritas interview, you described the reason why you blew the whistle on Facebook’s censorship as “people deserve to know.” Could you tell us about the very first time you found any bias among Facebook itself or content moderators working at Cognizant?

Hartwig: Yes. One of the first examples I saw was in the summer of 2018. There were instructions from Facebook to delete, to remove a viral video that showed a Trump supporter being victimized, being attacked. In that video, the Trump supporter was attacked in a restaurant and the attacker called the kid a bad name. There was not a good policy rationale for deleting this video. There wasn’t any nudity. It was a very minor ― I don’t even know if it was a violation. So that’s the first example of bias that I saw.

And then I noticed about 19 other examples in that first year. So that was before I even talked to Project Veritas. I wrote a letter to a few congressmen in May of 2019, but I did not hear back from them initially. After the Project Veritas video came out, I did communicate with Congressman Matt Gates from Florida and we did send a criminal referral to the Department of Justice. So initially, when I reached out, I did not hear back. But after I filmed with Project Veritas, then the congressmen and the senators were happier to talk to me.

I have a list here of about 40 examples total. And I am publishing a book in a couple of months that has the full list of examples of bias. So one good example is Facebook will make newsworthy exceptions to their own policy. So, for example, there is a national news anchor named Don Lemon, and on national television, he said that white males are terror threats. And Facebook even acknowledged, they told us that that phrase violates their hate speech policy. And so they said we are going to make a newsworthy exception to allow that phrase. So we see this double standard where they make exceptions for people of certain viewpoints, people on the left, but they do not provide that same protection to individuals on the right of the political spectrum. We also see their stances regarding abortion where they do not consider abortion as a violent death per their policy.

There are a few other examples here. This one is about the senator from Australia. He is a right-wing senator, and a kid broke an egg on the back of his head. He turned around and slapped the kid a couple of times, and that violated the child abuse policy. But Facebook allowed that to stay on the platform, and showed that right-wing senator being humiliated. It’s okay to show him being humiliated even if there is child abuse.

 

A Meme that Showed President Trump Shooting Himself Was Allowed

――Regarding your documents, a brutal picture which depicted President Trump shooting himself was NOT deleted by Facebook’s content moderator, since it “depicts self-harm,” despite Facebook supposedly holding out a policy to remove any posts that wish or hope for death. How can this kind of extremely one-sided decision go unpunished? Do you think Facebook didn’t know what its moderators actually did?

Hartwig: This is important to talk about because we do have a policy about suicide. It’s crucial that children and young people are not exposed to memes about suicide. Unfortunately, it was something we saw a lot; we had to see videos for example of people slitting their wrists, hurting themselves. So it’s very troubling.

However, despite such a policy forbidding depiction of suicide, there was a cartoon meme that showed President Trump shooting himself, and the guidance was that it was allowed. That was possibly in early 2019 when it was allowed, and since then, I believe they did tighten up their policy to disallow any form of a meme that would show Trump shooting himself. But for a long time, they did allow a lot of attacks against Trump. Whereas any time Trump would give a speech, Facebook would warn us. As content moderators, they would tell us to look for hate speech every time he spoke.

Also, there was another meme that showed a joke that talked about Ronald Reagan’s shooter. So President Ronald Reagan was almost assassinated, and that person who tried to assassinate the President spent many years in a mental health hospital. He was released a couple of years ago, and there was a meme that insinuated that that individual should kill Trump, and we were told to not delete that meme. Facebook told us not to delete that meme.

 

Leaving Hate Towards Whites, Police and Conservatives

――It is revealed that Facebook actually allowed hate speech toward those people on the black list. For instance, regarding a white police officer. Have you witnessed or heard any evidence of this bias?

Hartwig: Yes. I’ll give you the example of attacks on cops. Cops should be treated as private individuals. However Facebook modified their policy to allow more attacks on cops. And so, for example, if I call someone a pig, if I post a picture of someone and I call them a pig, that is called comparison to animal. It doesn’t matter who reports it and that post will get deleted. Facebook said that policy no longer applies to protect cops. So if I post a photo of a cop, and I say that cop is a pig, then Facebook said it stays on the platform.

――That is clear discrimination against police officers regardless of color.

Hartwig: Right. If Facebook’s whole goal is to prevent real-world violence, then why would they allow more attacks on cops? Because that could lead to cops being killed and we’ve had cops killed just because of hatred towards cops. It really goes against their stated purpose of preventing real-world harm, and that is one of the few things they should do is to try to prevent real-world harm. They are obviously not interested in protecting police officers.

They had this list of hate figures. During the time I was there, from what I remember, I think the only time they put someone on the left as a hate figure was, I believe, Louis Farrakhan. They added him to the hate figure list. But generally, the only people who violated or who are considered hate figures were – there were many White nationalists from that list – but I didn’t see any other types of nationalists that were placed on that list. For example, there’s a black nationalist named Marcus Garvey who advocated for segregation. And he was not placed on that list. It seems like there tends to be a disproportionate amount of white individuals placed on the hate figure list. I totally understand that there has been racism in our country, and we need to have equal rights. But there is a thing called reverse racism. And so it’s important to be pretty fair across the board.

There are a few other examples. In South America, there were LGBT movements, there were also movements that are pro-choice in favor of more abortions, and Facebook allowed topless protests. So if you’re protesting in the streets in Brazil or in Chile, then you can walk around, as a female, topless without violating a policy about nudity.

On the other hand, in the northeastern United States, there was a straight pride march. And Facebook saw this and they gave us guidance to watch out for hate speech coming from the straight pride march. So automatically, they just assume that straight people are haters towards LGBT, which is not the case. Also, the phrase “white trash,” initially, when I started working there was a violation of the hate speech policy. But then they reversed that. So now if I call someone white trash, it is no longer a violation.

Another great example is, in the summer of 2018, Facebook gave us an update about Pride Month. Every June is Pride Month in the United States, and it celebrates LGBT rights, which is great. So the thing is, Facebook specifically said it’s okay to attack straight white males as filth for not supporting LGBT. So attacks are not allowed against any other group except for straight white males. So why are they singling out one group and allowing more hatred towards one group?

 

Artificial Intelligence Was Used to Protected Greta Thunberg

――Before outsourcing censorship to allied companies like Cognizant, Facebook itself decides which post to ask a review of using their artificial-intelligence system. And it is said that this AI has a strong tendency to pick conservative posts as hate speech or fake news. Did you find this kind of bias when you were working as a content moderator?

Hartwig: Yes, Facebook has artificial intelligence, and the best example of this is with Greta Thunberg. About a year ago in December of 2019, there were many attacks on Greta Thunberg, the climate activist from Sweden. And so people were calling her retarded or they would call her “Gretarded,” which is a play on the words, and so Facebook said to delete all instances of that. At that time, Greta Thunberg was minor, but she was public figure of the year. And normally, you can call a public figure who’s a minor retarded. That is allowed. So once again, Facebook made an exception to their own rules to protect this individual who is a left wing climate activist. But yes, the AI helped us scour through and find these posts. So for an entire week, we were deleting anything related to Greta Thunberg that was an attack.

 

Coworkers Small Talk About ‘Assassinating Trump’

――May I ask about your ex-colleagues? I saw some secret filming which exposed how strongly left-biased Facebook’s content moderators were. Could you share with us any of their comments or behaviors representing their extremely leftist tendency or their anti-conservative bias? Have you experienced any shocking incidents inside Cognizant?

Hartwig: Yes. There was a coworker of mine named Cassie, and she was openly talking about assassinating Donald Trump. As far as I know, she did not get in trouble at work for that. She was talking about accepting the 80 million dollar bounty from Iran. This is a bounty given in January of this year(2020), almost a year ago. And so she said that she would take one for the team. That’s one example of that. And, then, another coworker, his name was Israel, I have him on camera talking about how much he hates Trump supporters.

We even have content moderators who were deleting anything mentioning MAGA, but they got away with doing that and were not punished.

Also when I first started the job in March of 2018, my trainer, her name was Perla, she was talking about how much she loves Obama. So it was a weird culture because you could talk openly about praising and supporting Obama, but you could not do that as easily talking about Trump.

This is a really good example. I had a co-worker named Saadi; he was on the policy team. He had a supervisory position and his profile picture in our work chat was of President Obama. I remember another co-worker changed his profile picture to Trump. But he only left it there for a couple of days, so it was okay to have a profile picture of Obama for two years. But none of my other co-workers had a profile picture of Trump, so I thought that was interesting.

I think it’s important to know that because Facebook’s headquartered in San Francisco, the culture in San Francisco is just very different, and it’s very one-sided towards liberals and leftist viewpoints, so it’s very hard to fight against that. If 80% of your employees are leftists, then it’s called group think, and it creates an echo chamber. So I think everyone knows that Facebook is leftist. And for years, many conservatives have been censored. Maybe inadvertently because of the group think. Perhaps they even don’t know how biased they are.

I do remember a conversation I had with a coworker, and I asked her – we were kind of whispering – and I said, “Do you think Facebook is liberal?” and she said, “Yes. Facebook is liberal as f***.” I apologize for the swear word. I’m just quoting her. And so I think secretly, regardless if they’re a liberal or conservative at work, they knew that Facebook was liberal and that influenced those policies. There was a tendency to allow attacks on Trump.

They did have a quality assurance team. And we were encouraged to think with our head, not our heart. However, it doesn’t matter; the most important thing to remember is the policy itself was biased. So the policy is designed by roughly six people on the Facebook policy team, and it’s a global policy team, so they make decisions that affect everyone around the world.

――Why do you think Facebook decided to outsource censorship to the company which is filled with left-biased employees?

Hartwig: One argument is that it would maybe distance themselves from legal action because they could argue that Cognizance, the company, was making their own policy decisions. I think that question has been asked by members of Congress when they had hearings with Mark Zuckerberg.

But remember that not only content moderators are biased, but Facebook’s policy itself is biased. Even if Facebook hired a thousand workers in Texas, which is a conservative region, if they have to follow Facebook’s policy, which is left, and if they’re good workers, they’ll be obedient, and they’ll follow the policy. So it doesn’t matter what the workers’ views are because if the policy is biased, then these workers are going to follow the policy.

 

Facebook Interferes on Elections Abroad

――At your Project Veritas interview, you pointed out Facebook’s content moderators were interfering on a global level in elections. How can they interfere on elections abroad?

Hartwig: Globally, Facebook has over 3.7 billion active users. This is a tremendous power that they have. And now they own WhatsApp, and Instagram as well. Facebook is extremely powerful. They control elections.

When I first got hired, I was a bilingual content moderator and was moderating content in Latin America in Spanish. In 2018, we had the Mexican presidential election, and Facebook gave us election training decks and showed us content about different candidates. We were told to treat all presidential candidates equally. However, Facebook made an exception to not allow attacks against the son of one candidate, Lopez Obrador. He is leftist, and there were attacks on his son. They were making fun of his hairstyle. And Facebook did not allow attacks on this kid’s hairstyle. That’s just one example.

 

Big Tech’s Censorship Is the Next Existential Threat to our Society

――You penned an open letter to Congress and urged lawmakers to limit Facebook’s power to stifle political expression.

Hartwig: I and two other whistleblowers, Zach Vorhies, the Google whistleblower, and also Zachary McElroy, another Facebook insider with Project Veritas sent an open letter to Congress talking about censorship. I think it’s important to have senators and congressmen talk to people like me who have experience and know about Big Tech and help them ask the right questions.

In addition to Facebook, Jack Dorsey, the CEO of Twitter testified recently and he said that Twitter does not have the ability to influence elections. I find that incredible. We need to take more action. There are congressmen in the United States who are taking action. Here, in Arizona, we have Paul Gosar who is fighting against Big Tech. And we have also Andy Biggs from Arizona who is doing a good job and Senator Josh Hawley, and Senator Ted Cruz as well. We need more people like them, because this is the next existential threat to our society.

The good thing is we were able to do a criminal referral to the Department of Justice. In June of 2020, the video went public with Project Veritas and I gave evidence to Congressman Matt Gaetz. Then he sent a letter to the DOJ for the criminal referral for Mark Zuckerberg, because in April of 2018, Mark Zuckerberg testified that Facebook does not censor political speech. But my evidence shows that they do censor political speech. So there’s a possibility that there’s alleged perjury from what Mark Zuckerberg said.

――So, we might be able to see progress in the process to stop Big Tech’s censorship. Thank you so much for sparing your time.