Mark Zuckerberg has likened fact-checking to censorship and decided to end Facebook’s program involving professional experts specifically tasked with verifying facts. Ilja Ber, founder of the outlet Проверено (Verified), believes the real issue lies not with fact-checkers but with Facebook’s algorithms.
“Eliminating fact-checking will further pollute the platform with the largest user base on the internet, filling it with even more disinformation. This will pave the way for new crises fueled by deepfakes, fabricated quotes, and conspiracy theories,” warns Ber.
Ber is no stranger to Facebook’s inner workings. Since 2022, he has been participating in Facebook’s fact-checking program in collaboration with the outlet Delfi. Drawing from his extensive experience, he offers a critical perspective on Zuckerberg’s recent pivot, shedding light on the potential consequences of abandoning the fact-checking framework.
Fact-checker Ilja Ber lives and works in Tallinn and has been declared a foreign agent by the Russian Ministry of Justice.
Pressure Forced Zuckerberg to Hire Fact-Checkers
Facebook launched its independent fact-checking program in 2016 under intense public pressure that followed Donald Trump‘s unexpected victory in the U.S. presidential election. At the time, there was serious consideration of the possibility that a disinformation campaign on social media, allegedly orchestrated by Russia, had decisively influenced the election results. As a response, Facebook decided to involve third-party organizations—independent fact-checking projects certified by the International Fact-Checking Network (IFCN)—to verify questionable posts. These experts assessed the reliability of posts and flagged identified fake news with a special label.
On January 7 of this year, following Trump’s second victory in the presidential election, Facebook founder Mark Zuckerberg announced the end of the fact-checking program. Initially, the decision affects only Facebook’s operations in the U.S., where contracts with 10 fact-checking organizations—covering content in more than 60 languages—will be terminated. However, it is not ruled out that Zuckerberg may also discontinue independent fact-checking elsewhere.
“It’s time to return to the roots of free speech,” the billionaire explained his decision. He referred to the former fact-checking partners as “politically biased censors” and accused them of fostering distrust within American society.
This time, it is Trump’s team that is pressuring Meta to firmly ensure “freedom of speech.” Once again, Zuckerberg has succumbed to pressure—though this time, it’s from the opposite direction. Below, I will analyze the claims Zuckerberg used to justify his decision.
Fact-checking is censorship, and fact-checkers are censors.
The Estonian dictionary Sõnaveeb defines censorship as: “Censorship carried out by authorities, primarily in totalitarian states, to prevent the spread of undesirable information or ideas.” Thus, the primary task of censors has been and continues to be combating the dissemination of information unfavourable to those in power.
Fact-checkers, however, deal with something entirely different. They do not remove the information but instead, add it. “They enrich public debate by providing context and facts, enabling every responsible citizen to form their own opinion,” explained the European Fact-Checking Standards Network (EFCSN) in response to Zuckerberg’s criticism.
Moreover, fact-checkers preserve fakes for reference in case they are removed from the web. They save web pages, take screenshots of social media posts, and download videos for archival purposes.
In the world of information, fact-checkers perform the same role for information as doctors do for people. They examine all the circumstances and provide their assessment—essentially, they make a diagnosis.
The methodology of professional fact-checkers requires every claim to be substantiated with references to reliable sources. Furthermore, fact-checkers themselves do not engage in “treatment”—they do not remove flagged posts or restrict user accounts. The fate of flagged posts is determined by the moderators and algorithms of social media platforms, operating under the rules established by the platform’s leadership.
On Facebook, which has three billion users, the role of algorithms in fact-checking has gradually increased because the sheer scale of content surpasses the efficiency of human oversight. Meta, the parent company of Facebook and Instagram, states on its website:
“Our review technologies detect and remove a significant portion of rule-violating content even before anyone reports it. We remove millions of rule-violating posts and accounts from Facebook and Instagram every day. Sometimes, our system identifies potentially rule-violating content but refers it for human review and action. This work never stops.”
However, the more decision-making was entrusted to algorithms, the worse the outcomes became. The number of false positives increased, user dissatisfaction grew, and the platform began to lose its reputation as a place for free and safe self-expression.
It is important to emphasize that Facebook rarely fully blocked users for systematically spreading misleading information. There were few exceptions, but they were often controversial. For example, some accounts of Trump supporters were blocked, including Trump’s own account, which was suspended in January 2021 for supporting the rioting crowd that stormed the Capitol.
Most of Facebook’s bans were related to violations of principles prohibiting the dissemination of content such as nudity or sexualized material, incitement of hatred, posing real threats or causing real attacks against individuals or groups, using fake profiles, and so on.
At the same time, a large amount of content, including advertisements, currently spreads on Facebook that blatantly violates these principles. The system functions imperfectly and will continue to do so—now without fact-checkers in the U.S. Essentially, Zuckerberg, who likely understands exactly how his social network operates and who is responsible for what, decided to scapegoat the fact-checkers.
Fact-checkers are politically biased
This is an unfounded claim that is not supported by evidence—because none exists. The only somewhat notable figures have been presented in a disinformation study by Harvard University’s Kennedy School. The authors of the study surveyed 150 fact-checkers from different countries. The study states:
“Experts largely leaned toward left-leaning political views: far-right (0 individuals), moderately right-leaning (0 individuals), moderately right-centrist (7 individuals), centrists (15 individuals), moderately left-centrist (43 individuals), moderately left-leaning (62 individuals), far-left (21 individuals).”
Source: Screenshot from IFCN.
These responses were likely based on the self-identification of the surveyed experts. This is insufficient to draw such categorical conclusions as Zuckerberg has. However, I must admit that my intuition and personal experience suggest that a “left-liberal bias” may indeed exist among fact-checkers.
What matters more is something else. Personal political views are something everyone has, regardless of their profession. If a person strives for accuracy and adheres to the code of conduct for fact-checkers, their views will not influence their ability to distinguish falsehoods from truthful information.
Fact-checking focuses on the most virally spread claims and facts suspected of being false. There are no known cases where fact-checkers have faced specific accusations of neglecting to verify claims about a particular politician for political reasons—or vice versa. Instead, abstract accusations are made, when the process should work the other way around: first come the facts and studies, then the conclusions.
Fact-checking is being replaced with “Community Notes”…
… “just like on X, where they work excellently.” Zuckerberg announced such a change, but the new tool has not yet been implemented, and its details remain unknown. However, it is instructive to examine how “Community Notes” work on X. Users of the platform can add notes to controversial posts, providing context and explaining possible inaccuracies by the author. These notes are first evaluated by other community members and only appear under the post once they have gained sufficient support.
There is no verification of the participants’ expertise or impartiality in the community. The system is poorly protected against manipulators who might be paid to achieve specific political or commercial goals. Such cases have been documented on platforms like Wikipedia.
The system has allowed serious errors with significant consequences. For instance, in October 2023, claims surfaced that a photo posted on Israeli Prime Minister Benjamin Netanyahu’s official X account—showing a child allegedly killed by Hamas fighters—was fake. A “Community Notes” annotation appeared under a repost of the photo by political commentator Ben Shapiro, suggesting it was a forgery. However, it later emerged that the photo, which caused a major scandal and significant reputational damage to the Israeli Prime Minister, was genuine.
Last May, X announced that approximately half a million users worldwide had registered to contribute to “Community Notes.” However, X does not disclose how many of these individuals are actually active. For comparison, in July last year, 325,000 people were registered as editors for the English-language Wikipedia, but only 38,000 of them actively participated in writing new articles and editing old ones on a regular basis.
Wikipedia is often cited as an example of successful crowdsourcing—effective voluntary collaboration among people. Based on this experience, some suggest that “Community Notes” could also function effectively. However, this overlooks a key distinction: Wikipedia is a secondary source of information, and its content must be based on reliable primary sources. When it comes to newly emerging fake news, reliable primary sources are often nowhere to be found.
Is fact-checking ineffective?
Source: Screenshot.
What can be said about Zuckerberg’s claim that fact-checkers “destroyed more trust than they created”? As recently as 2024, before the European Parliament election campaign, Meta praised the effectiveness of its labelling system: “From July to December 2023, over 68 million pieces of material shown to residents of European Union countries on Facebook and Instagram received fact-checker labels. When a warning label is added to a post, 95% of people do not click on it and simply scroll past.”
What does the elimination of fact-checking mean for regular Facebook users?
It is unlikely that there will be more fake news, as even now, very few people hold back from posting, knowing there is little fear of being banned or having their account closed for violating Facebook’s policies. On the other hand, the impact of fake news may increase because it will appear more frequently and without labels in news feeds, with most of it going unmarked. This raises the likelihood of major scandals and crises caused by deepfakes, falsely attributed quotes, and the spread of conspiracy theories.
One way or another, Mark Zuckerberg has legitimized the populist claim that fact-checkers are censors. According to him, all that needs to be done is to “get rid of” them (as he explicitly said), and his platform will return to a blissful era.
Ilja Ber’s article was originally published in Insider.
