By almost any measure, the first three months of 2021 were the worst ever for online freedom. Amazon, Twitter, Apple, Google, Facebook, YouTube and others proved to the world that the Big Tech censorship of conservatives is a reality. And they did so in disturbing, authoritarian ways that highlight their unchecked power over information and our political process.
At least 10 separate tech platforms silenced then-sitting President of the United States Donald Trump over the speech he gave in Washington, D.C. the day of the Capitol riot. Both Google and Apple pulled Parler from their app stores. Amazon Web Services also cancelled its contract with Parler, shutting the burgeoning social media site down for more than five weeks. Google removed LifeSiteNews from its advertising programs, and YouTube shut down its channel. YouTube demonetized Steven Crowder’s channels. Amazon removed Ryan T. Anderson’s book examining transgender ideology, and it also removed a Clarence Thomas documentary while continuing to sell at least 270 hate items.
That consistent assault on free speech makes 2021 an ideal time to track the biases and failures of Big Tech. Starting with this quarter and going forward every three months, the Media Research Center’s Free Speech America operation will grade the top Big Tech companies, including social media, search media and others that form the backbone of our online lives.
This quarter, Big Tech earned a collective “F.” Every one of the Big Tech companies reviewed got an “F” in free speech. That’s simply appalling. The quarter began with unparalleled restriction of the president of the United States. It was unquestionably the low point for online freedom since the creation of the internet.
However, the quarter ended with Facebook releasing new tools that gave users a bit more control over the content they choose to see in their feeds — a small win for users. Facebook is in the midst of its appeals process that will determine whether Trump will be allowed back on its platform, despite aggressive lobbying by the left and major media.
Those were not big steps, but if social media companies want to win back the trust of their users, such reforms are a start. More freedom and less restriction is the only model that celebrates American principles and treats users as adults who can make their own decisions.
The fight, like all major fights now it seems, comes down to safety vs. freedom. The far left, including Big Tech, values the fantasy of an online safe space over the very freedoms that gave rise to their products.
The Free Speech America grading system includes grades in five categories culminating in an overall grade for each company’s Online Freedom. The five areas the companies were rated in include: Free Speech, User Transparency, Bias, Responsiveness to User Complaints and Fact-Checking.
Overall Grade: F
Amazon was another platform that pulled the plug on Parler using unsupported claims. Parler used Amazon Web Services (AWS) to power its website, so when Amazon cancelled its contract with Parler, the site was shut down entirely for around five weeks until it found a new host. Amazon also removed a book that questioned transgender ideology, later claiming that it would not sell books that claimed transgenderism is a mental illness. The author of the aforementioned book explained that his book did not make such a claim. The removal of the trasgender book itself marked a change to Amazon’s previous policy to not ban books containing hate speech, a change that it quietly debuted. It remains unclear why Amazon removed a documentary about U.S. Supreme Court Justice Clarence Thomas during Black History Month.
Free Speech: In the first quarter of 2021, Amazon took multiple actions against conservative content. Despite a contract for service with Parler, Amazon booted Parler from its Amazon Web Services platform, forcing Parler to temporarily shut down. Amazon supported its action with debatable claims that Parler allowed content organizing the Capitol riot on Jan. 6, though only eight of the 226 Department of Justice charges cited Parler. In the middle of Black History Month, Amazon removed the documentary “Created Equal: Clarence Thomas in His Own Words” from its streaming service. Amazon also removed Ryan T. Anderson’s book “When Harry Became Sally: Responding to the Transgender Moment.” Grade: F
User Transparency: After some pressure, Amazon finally gave a reason for removing Anderson’s book, saying that it would “not sell books that frame LGBTQ+ identity as a mental illness.” However, Anderson responded that “[n]owhere have I ever said or framed LGBTQ+ identity as a mental illness.” No explanation has been provided for why the Justice Thomas documentary was removed. According to Buzzfeed, Amazon told Parler that it was terminating its service due to “a steady increase in this violent content on your website, all of which violates our terms.” Amazon was not clear about how much its decision was influenced by its own employees from a group called “Amazon Employees for Climate Justice,” who had demanded that the platform remove Parler. Grade: D
Bias: While Amazon took action against several prominent conservative products, it was found to be selling at least 270 hateful items. At the time the study was done, efforts to find similar items promoting violence toward Democrats was fruitless. Though it pulled Anderson’s book taking a stance against transgender ideology, it did not pull a book responding to Anderson’s book called “Let Harry Become Sally.” It also sold books such as Hitler’s “Mein Kampf,” Mussolini’s “My Autobiography,” Marx’s “Communist Manifesto,” and the Unabomber’s manifesto despite saying that its new policy prevented selling any books containing “hate speech.” It pulled the Clarence Thomas documentary during Black History Month while continuing to stream documentaries about radical leftists. Grade: F
Responsiveness to User Complaints: Amazon stated that it has used “a combination of machine learning, automation, and dedicated teams of human reviewers'' in order to automatically ban items such as the conseravtive ones listed above, but it is very slow to reverse these decisions once incorrect decisions are brought to its attention. While some of the hateful items found on Amazon’s site have been removed, many were not. Overall Amazon was not responsive to user concerns or questions regarding censorship or bias. Grade: F
Overall Grade: F
Twitter has consistently been one of the worst platforms for censorship, constituting the majority of entries in the Media Research Center’s CensorTrack.org database. In fact, its roughly 250 entries in the first quarter more than doubled the next closest platform. This excessively terrible showing in the Free Speech category brought its overall grade down to a resounding “F.” Twitter was among at least nine other platforms that dumped the president’s account. The censorship was a reversal of a policy over 18 months old, in which Twitter maintained that it was more important to allow global leaders the platform to speak than to censor what they say. For that time, Twitter applied warning labels on Trump tweets that it deemed inappropriate. Twitter acknowledged that this amounts to censorship in a blog post last year, saying “[t]weets with labels are already de-amplified through our own recommendation systems.” Twitter later stated that the Trump ban was unequivocally permanent. Twitter also removed the Project Veritas account and censored other prominent conservative accounts like Dan Bongino and Catholic World Report.
Free Speech: Twitter has dominated the entries in the CensorTrack database. This quarter, Twitter banned then-President Donald Trump, and said it would not reverse the decision. It censored Steven Crowder on at least three occasions this quarter. It permanently removed the Project Veritas account and censored its founder James O’Keefe at least three times. Twitter also censored conservative radio host Dan Bongino, Sen. Marsha Blackburn (R - TN), Catholic World Report and Rep. Marjorie Taylor Greene (R-GA) in this quarter as well. Grade: F
User Transparency: Twitter often does not provide a reason for taking action against a user, or does not provide information as to which specific tweets were in violation. At other times, the explanation provided does not make sense, and no clarification is provided. Twitter did release some specific guidelines on what would lead to a suspension over COVID-19 content this quarter, a small step toward more transparency. Its other guidelines are still vague and subjective. Grade: D
Responsiveness to User Complaints: While suspensions can be appealed, the appeals often take far longer to resolve than the suspension would have lasted. Thus, the lengthiness of the appeals process has led many users to not even file an appeal, or to abandon their appeals while waiting for a response just so they can get back onto Twitter. Twitter has reversed some of its decisions, typically calling them “errors” if it was a conservative account, though no leftists being suspended in error have been observed. Grade: D
Fact-Checks: Twitter rarely applies a fact-check label to a tweet. When it has, it usually just added a link below the tweet, not an interstitial filter that would greatly reduce engagement and reach. It used such notices prolifically this quarter in relation to tweets about the election or about COVID-19. When used, the fact-check pages that are linked to are not usually one “authoritative” article written by a fact-checking organization as Facebook’s are. Rather, Twitter has typically created a page that is a roundup of many articles and tweets about the topic, though nearly all of them come from left-leaning sources. Grade: D
Overall Grade: F
Apple is unlike the other platforms on this report card. At present, it only had the opportunity for censorship in its App Store and its music and book stores. However, the removal of the Parler app from its App Store alone was an egregious enough violation that it made the list for this quarter.
Free Speech: This quarter, Apple had one huge incident when it removed the Parler app from its App Store. The app was removed despite the hypocritical narrative used to remove it being questionable. Beyond that, Apple also briefly removed the Ryan T. Anderson ebook, “When Harry Became Sally: Responding to the Transgender Moment.” That decision was quickly reversed. Grade: F
User Transparency: Ryan T. Anderson found out that his book had been removed by Apple Books from a Twitter user. That user also later reported that it had been restored. Apple never informed Anderson of the removal, and no reason for the temporary removal was offered. The company was fairly transparent with its policies. Most apps would not be removed from the App Store because apps are vetted before being added. Grade: D
Bias: Similar to Google, Apple pulled the Parler app over claims that the Capitol riot was organized unfettered on the Parler platform. However, Apple’s take doesn’t tell the full story. It was reported that Facebook was the source of much of the organizing. Despite the reports on Facebook’s involvement, Apple did not even threaten Facebook to clean up its platform. Apple CEO Tim Cook said that the App Store will only reevaluate the Parler app once it has “put in the moderation that’s required to be on the store,” ignoring the fact that Parler has had a very transparent moderation policy. Ryan T. Anderson’s transgender book was briefly removed, but quickly restored. Apple has not removed leftist radical books such as “Mein Kampf,” the “Communist Manifesto,” or Mussolini’s “My Autobiography.” Grade: F
Responsiveness to User Complaints: Apple quickly reversed the removal of Anderson’s book, but reportedly won’t consider restoring Parler unless it adds a moderation policy that Apple approves of. Parler had a moderation policy in place when Apple pulled the app, though seemingly not one to Apple’s standards. The platform was somewhat responsive overall to user complaints and questions in this quarter, but better responsiveness with more specific guidelines on what is required to change in order to be restored to the App Store is needed. Saying that Parler must “put in the moderation that’s required to be on the store” was far too vague when Parler already had a moderation policy that appeared to be more transparent than Facebook, neither of which were even threatened with removal from Apple’s App Store. Grade: D
Overall Grade: D
Google has not censored in the same way that Facebook and Twitter have, but it found plenty of opportunities to exert its control over content that it didn’t like this quarter. Its algorithms control content in massive and unseen ways, making its bias more difficult to show. Google removed Parler from its app store, and it removed LifeSiteNews from its advertising programs. Google was called out for suppressing search results for Blue Anon. It also utilized Wikipedia content in its search results, a heavily biased and often unreliable source.
Free Speech: Google was heavy on obvious censorship this quarter. It first removed the Parler app from its Google Play app store. A class-action lawsuit was filed against Google, alleging its Play store is a monopoly. Google cited “user safety” after the Jan. 6 Capitol riot in removing Parler, though it has since been shown that the narrative of Parler being used to incite violence was questionable. It then threatened similar action against another free speech-minded social network, Minds. Minds instead made some changes to its app to avoid getting deplatformed, which has made the app nearly unusable according to many recent reviews. Google removed LifeSiteNews from all of its advertising programs, citing “alleged ‘dangerous or derogatory content.’” Google also censored nearly 100,000 negative reviews of the Robinhood app when it froze trading of certain stocks. Grade: F
User Transparency: While Google typically provides reasoning for its actions even when the reasoning is unfounded, LifeSiteNews said Google did not explain what content was considered “dangerous or derogatory.” When deplatforming or demonetizing sites, the complete reasoning must be provided or it may be seen as arbitrary or ideologically based bias. Grade: D
Bias: Google pulled the Parler app, hypocritically claiming Parler did not do enough about posts calling for violence. However, Parler “warned the FBI more than 50 times before the Capitol riot,” according to ARS Technica. In fact, the left agreed that Facebook hosted much of the organizing, yet no similar action against the Facebook app was even hinted at. As of publication, a Google search for "Blue Anon" showed that results are still mostly being suppressed a month after such suppression was first noticed.Grade: F
Responsiveness to User Complaints: Google was not very responsive to questions and complaints. Despite being called out for bad Blue Anon search results a month ago, nothing has changed. It has also not taken action to reverse its decision on the Parler app, despite evidence coming out that Parler reported troubling content to the FBI in advance of the Capitol riot. Grade: D
Fact-Checks: While Google does not use fact-checks like Facebook, it does rely heavily on Wikipedia for content that is provided in “knowledge boxes” above search results. A study found that only one-third of Google searches result in clicks, in part due to the knowledge boxes that seemingly provide the information sought. However, Wikipedia’s co-founder Larry Sanger has called out the platform for being heavily biased. It is also subject to pranks and trolls, one of which resulted in Google showing Wikipedia content listing Trump’s former acting Director of National Intelligence Richard Grenell as President of the United States during this quarter, an embarrassing episode for Google. Grade: D
Overall Grade: D
Facebook was one of at least 10 platforms that removed Trump, setting the tone for a terrible quarter. Even its heavily leftist Oversight Board overturned Facebook’s censorship in four of the first five cases it ruled on. That’s a sign of how skewed the site’s content restrictions have been. Facebook censored at least 20 high profile accounts, including the late Rush Limbaugh. Facebook Spokesman Andy Stone apologized profusely for a singular case of high profile censorship of a leftist. Facebook’s fact-checking apparatus is anything but unbiased, and has produced horrible results. Facebook itself has admitted that posts with fact-checks on them receive 95 percent less engagement, regardless of the veracity or even relevance of the fact-check applied.
Free Speech: Facebook censored then-president of the United States Donald Trump, and it finished the quarter by removing a video of his daughter-in-law Lara Trump interviewing the former president. Facebook claimed it did so because the president’s ban was a ban of his voice, not just his account. It also produced at least 20 high profile cases of censorship including prominent conservatives such as actor Kevin Sorbo, independent journalist Tim Pool, actor Nick Searcy, former One America News Network host Liz Wheeler, conservative YouTuber Mark Dice, former Congressman (R-TX) Ron Paul, The Blaze journalist Elijah Schaffer and the late conservative radio icon Rush Limbaugh. Grade: F
User Transparency: Facebook’s Oversight Board is badly skewed away from Americans to rule on American content. However, it has so far mostly supported free speech, despite disturbing comments from several members of the board. In addition, Facebook ended the quarter offering users new tools that allegedly will allow users to overrule some of the authoritarian impulses of the worthless fact-checking system. Facebook often fails to tell users why their accounts have been censored, or specifically which posts caused the infraction, even when asked. Grade: C
Bias: Facebook regularly removed and suspended accounts from the right, but rarely from the left. Even when leftist Congresswoman Marie Newman (D-IL) was censored over a video of her placing a transgender flag outside her congressional office, Facebook spokesman Andy Stone was quick to have the censorship reversed and issued a profuse apology. In a rare move against the left, Facebook did remove Louis Farrakhan’s vaccine misinformation. Grade: D
Responsiveness to User Complaints: Facebook is horribly bureaucratic and slow or even immobile on many complaints. It does address some user concerns, and some unfairly removed content was restored. But it takes far too much time and there is no effective process to do so. The Oversight Board rulings took over a month to decide, and even longer for the Board to agree to hear them. In particular, the Oversight Board has still not resolved whether to allow Trump back on the platform. Grade: D
Fact-Checks: One of the most frustrating aspects of the Facebook user experience is the biased parade of fact-checkers intent on enforcing a leftist interpretation of events. There are many examples where Facebook applied a fact-check that cited an article that has absolutely nothing to do with the post whatsoever. Some off-the-mark fact-checks included fact-checks under posts by The Blaze commentator Steve Deace and conservative radio host Todd Starnes promoting their respective Parler pages and various Bill O’Reilly videos posted on The Resurgent’s Facebook page. Facebook also fact-checked a Liz Wheeler video discussing a bill introduced in New York, labeling it as “partly false” because the bill was not yet scheduled for committee consideration and did not have a co-sponsor. Facebook’s fact-checking organizations arbitrarily came after a Dr. Seuss meme, a Family Circus meme, and The Federalist. The fact-checking system is irredeemably horrible, and the concept of it is autocratic and condescending. Grade: F
Overall Grade: D
YouTube set the tone for its quarter by banning Trump along with at least nine other platforms. It followed this up by boasting in late January about having removed over 500,000 videos that violated its COVID-19 Medical Misinformation policy since Feb. 2020. YouTube continued the quarter with multiple other high profile cases of censorship, including completely deleting the LifeSiteNews channel and multiple copies of Trump’s speech at the Conservative Political Action Conference (CPAC). The platform also removed video from a Senate hearing about COVID-19.
Free Speech: This was a bad quarter for YouTube censorship. After banning then-U.S. President Donald Trump, the platform removed his CPAC speech from the American Conservative Union’s channel and from the Right Side Broadcasting Network’s (RSBN) channel. It suspended RSBN’s channel for two weeks after removing the video. It also demonetized The Epoch Times. YouTube deleted the LifeSiteNews channel for “putting out medical misinformation,” which the site vehemently denied. Video from a Senate hearing was removed from Sen. Ron Johnson’s (R-WI) channel but remained on other channels, and Dr. Drew Pinsky was deplatformed for a week. Steven Crowder was censored multiple times by YouTube this quarter. Grade: F
User Transparency: YouTube provided a clear answer as to why it was dumping Trump, but it has not been consistent with its censorship explanations. YouTube told Steven Crowder why he was being demonetized, even admitting that advertiser pressure played a part, according to Louder With Crowder. However, even when YouTube has cited a specific policy violation, in many cases it was unclear how the content violated the referenced policy. For example, Dr. Drew Pinsky was left guessing why his video was removed, as was political commentator Calvin Robinson. Users cannot be expected to stop violating policies when they are not told how their content has violated those policies. Lack of transparency is a slippery slope that too often ends in permanent punishment of the user. Grade: D
Bias: While an abundance of censorship of conservative content was catalogued, the only notable cases of leftist censorship centered on footage from the Jan. 6 Capitol riot, which was also removed from other accounts. There were reports that YouTube removed dislikes from videos on the Biden White House channel, which YouTube confirmed, though the platform claimed that they were spam. YouTube only reversed one notable instance of censorship this quarter, and while it said that it would reverse the Trump suspension when the risk of violence had passed, this is a subjective target and the account has yet to be restored. Grade: F
Responsiveness to User Complaints: YouTube has responded fairly well to complaints and questions. The platform is typically very responsive to users tweeting questions about suspensions. Even if the outcome was not favorable, the platform generally responded. YouTube’s CEO, Susan Wojcicki, said that it would restore Trump’s account when the risk of violence has passed. It also reversed a termination of the BBC-affiliated talkRADIO channel. Grade: C
Fact-Checks: YouTube introduced fact-checks in the U.S., much like what Facebook uses, in April of 2020. However, in recent months it has simply added labels below certain videos that point users to other sources of information about what is in the video. For example, many videos about COVID-19 that have not been removed have a notice below them that says: “Get the latest information from the CDC about COVID-19” and provides a link to the CDC’s COVID-19 web page. The CDC’s guidance throughout the COVID-19 pandemic has raised a lot of questions and concerns, so using it as the sole source of confirmation on COVID-related videos is questionable at best. Since YouTube boasted about removing over 500,000 COVID-related videos, using a questionable source for its determinations on what videos should be removed would not be advisable. Grade: D
PARLER, GAB and Others
When conservatives complained about the censorship of their voices by Big Tech platforms, a common response was that they should start their own platforms. Several new platforms have been created claiming to have done just that. Despite the fact that some of the more prominent alternatives, like Parler and Gab, have experienced their own censorship by the same Big Tech platforms in an attempt to shut down conservative alternatives, new platforms continue to pop up.
Because these platforms are so new, there was not much data on how these platforms will moderate content. Parler’s guidelines already explained that it does not allow certain kinds of speech, despite what the leftist narrative has pushed. It also outlined a community review system where a group of users who were trained on what is and is not acceptable must have a majority vote in order for content to be removed. Parler also instituted a point system, which it has been transparent about, that can ultimately lead to users being suspended or banned from the platform. It has not really been the “Wild West” that the leftist media would have you believe.
Given the newness of these alternative platforms, the new platforms were not evaluated in this report card. As more data become available on what sort of content these platforms are removing and how they are handling user complaints, the new platforms may be added to future report cards. New free speech oriented platforms that will be monitored include Parler, Gab, Rumble, MeWe, Clapper, Telegram, Locals and CloutHub.