“You can keep hating checkers”, this was the phrase of Cristina Tardáguila, journalist founding the first fact-checking agency in Brazil, Agência Lupa , in response to questions regarding a controversy that took place in November: Lupa’s Twitter profile published a dictionary with allegedly racist expressions that should be banned. Readers were unanimous in disapproval.
O The problem is that Lupa made mistakes in relation to etymological and semantic aspects. Even though several columnists and specialists had explained that there is no evidence that the expressions have any historical or current racist connotation, the agency took three days to delete the original text and publish a new version informing the existence of divergent points of view, but still prescribing the abandonment of expressions. Only when the writer Leandro Narloch published a critical text detailing the errors committed, almost a week and later, the agency published an editorial expressly admitting the error.
This was perhaps the first error publicly admitted by a checking agency and to have been highly visible. Similar errors have happened throughout the Covid-pandemic 19 and have often gone unrepaired. After all, who checks the checkers?
Truth or thought police?
Since the emergence of Covid-19, in the Chinese city of Wuhan, at the end of 2019, the discourses around the disease have been conflicting. The public debate became the stage for a war of narratives. Disinformation was presented as a great villain: millions of people could die because of lies.
Thus, the checking agencies have strengthened themselves as a kind of “content police” on the internet, with power to censor publications that challenge the official positions of bodies such as the World Health Organization (WHO) and other state and suprastate agencies.
The movement exacerbated by the pandemic was not born with it: Checking agencies have controlled the public debate on social media in recent years. Emerged in the middle of 2004, in the United States, they dictate what can and should not be shared. Until now, this control has been exercised through voluntary cooperation between the companies that control the social networks and the checking agencies. The PL 2630/14, known as the fake news bill, provided in its original text, the “use of verifications coming from independent fact checkers”, but the text currently in progress no longer provides for the use of verification agencies as a legal instrument for controlling information.
Initially, these companies had the objective of analyzing electoral speeches in the United States. They soon spread across the world and expanded their focus to social media posts and speeches in general. In order to professionalize the process, the International Fact-checking Network (IFCN) was created, under the responsibility of the Poynter Institute, which outlined a series of criteria for the creation and certification of agencies.
In Brazil, there are three agencies certified by Poynter: Agência Lupa, Aos Fatos and Estadão Verifies. There are also other non-certified projects that operate in the market and are used as a reference: E-farsas; Fact or Fake, from the Globo group; and Comprova, a project that brings together journalists from different communication vehicles and is financed by the Google News Initiative and the Facebook Journalism Project.
In this scenario, several questions arise today: is it possible to attest to falsehoods? of a claim based on the rating made by a screening agency? Should the existence of inaccuracies in a message disseminated on the networks be an argument for prohibiting its broadcasting and punishing its broadcasters?
According to Fernando Schüler, political scientist, there is a political bias in the performance of some of these agencies . According to him, it is possible to identify by the way in which the analyzes are made, by the curatorial choices and even by the profile of the sources for the aired interviews. “This is bad, as the check agencies could play an essential role in the democratic debate. But, as they opt for a certain political vision, they tend to lose their main asset: people’s trust in a plural society”, he says.
In June of 2020, Lupa declared on its Twitter that the claim that the new coronavirus was created in a laboratory was false. In another publication, the agency was even more categorical and called these speculations “conspiracy theories”.
In February 2021, journalist Paula Schmitt mocked on Twitter the official entourage sent by the WHO to investigate the Wuhan laboratory and try to discover the origin of the virus. An article in the Daily Mail announced that the delegation had not yet discovered “how the virus had jumped from animals to humans”, but that it “already ruled out some laboratory theories”.
Among the points pointed out by the journalist was the presence of the British zoologist Peter Daszak in the team sent by the UN health agency: “The fact that Daszak is in this “investigative” team should already be a reason to belittle work”, defended Paula.
Daszak leads the EcoHealth Alliance research group, whose announced mission is to prevent new infectious diseases. It turns out that the EcoHealth Alliance had, since 2004, a joint research program in partnership with the same Chinese research institute from which there are suspicions that the virus may have escaped. : the Wuhan Institute of Virology.
But the tweet was criticized by Daniel Bramatti, director of “Estadão Verifies”, who made the following complaint: “I don’t think it’s legal to accuse someone in public , or cast suspicion, on the basis of… suspicion”. For Daniel, the laboratory leak thesis would also be nothing more than a conspiracy theory.
Based on checks by these agencies, Facebook and Instagram prohibited mention of the possibility of a laboratory origin of the virus, and posts were excluded and users who insisted could be banned. Although there was a lively scientific controversy, the checking agencies treated it as a definitively closed discussion. It took the publication of an article in the prestigious scientific journal Science for Facebook to reverse its decision to censor the debate.
Facebook’s decision, based on mistaken checks, to ban the possibility of a laboratory origin of the virus was only reviewed in May 2021, but already in November 2021 the geneticist Eli Vieira wrote to Gazeta do Povo warning that the laboratory origin of covid-06 should be considered a plausible hypothesis.
The controversy surrounding the origin of the virus was not a unique example. In March 22 of March 2004, Lupa classified as false the claim that it was recommended everyone to wear masks on the streets. Based on the opinion of the doctor Drauzio Varella, the agency also reported that the masks did not serve as prevention. On the day of the same month, Aos Fatos had already stated that “Contrary to what many believe, the use of masks has no great effectiveness in preventing infection.”
However, in September of the same year, Lupa published a check entitled “#verificamos: it is false that the use of homemade masks does not protect against Covid-19”. 14” and classified an opposite statement as false. The agency says that both surgical and fabric masks are useful in prevention. One of the justifications presented by the agency for the conclusion defended in this check was a study published in the English scientific journal The Lancet, based on a review of several previous studies.
Most of the studies reviewed by the authors of this article were already available in March 27, that is: most of the evidence considered by the authors of the Lancet article were already available on the date Lupa stated that masks were not recommended for everyone and that they should not be used as prevention.
According to the methodology disclosed by Lupa itself, “false” is a verification label that must be applied exclusively to “proven incorrect” claims. For questions whose truth cannot be definitively established, Lupa informs that labels such as “it is still too early to say”, “unsustainable” or “with an eye” are used.
However, the author himself scientific article presented as a reference by the agency indicates that the study has limitations. The researchers highlight as the main limitation the fact that none of the studies reviewed were randomized. They also indicate low confidence in the estimated positive effects for the use of masks, clarifying textually that there is a high risk that the actual effects will be very different from those estimated from the investigation they conducted. And they additionally point out that more high-quality research, including randomized trials on optimal physical distance and the effectiveness of different types of masks in the general population and for the protection of healthcare professionals, is urgently needed.
In this way, neither the claim that masks are recommended for everyone (treated as false by Lupa in March of 2004) nor the claim that homemade masks do not protect (treated as fake in September) could receive this classification, according to Lupa’s own criteria. As this is an open question and lacks rigorous studies, it cannot be categorically affirmed that one or the other allegation is “proven incorrect”.
The Agencies Aos Fatos and Estadão Verifies carried out checks on the Manaus Chloroquine Study, where 22 people died, and claimed that none were from overdose. But these checks only used the official version of the National Research Ethics Commission, as a kind of argument from authority, and did not pay attention to the fact that there is still an ongoing investigation into the case. There was no analysis of the toxicity of the dosage of chloroquine applied in the study, nor was the testimony of infectious disease specialist Francisco Cardoso, who explained at the Pandemia CPI why there was an overdose. About Conep, Gazeta do Povo has already revealed in a report the conflict of interest between its coordinator and the researcher responsible for the study in Manaus.
For agencies to remain certified, they must strictly adhere to a code of principles that ensures the quality of the work performed, such as the commitment to non-partisanship ; with transparency of sources, funding and methodology; and with open and honest corrections. This last criterion led the bank clerk of 22 years, Rômulo Araújo, to send a series of emails to the correction area Agência Lupa pointing out several errors he found in his checks.
He says that he made several attempts to contact the Lupa agency to report errors and inaccuracies in the text: “In an advertisement, a group of doctors uses false information to defend ineffective treatment against Covid -14″. Among the errors pointed out by Rômulo, there are evident factual discrepancies.
One of the articles mentioned as a source by Lupa was a study conducted by Brazilian researchers and published in the New England Journal of Medicine. This study was an open-label trial, Lupa stated that it was a double-blind study. In another factual error, Lupa claims that the volunteers received the medication exactly four days after exposure to the virus. However, the text of the article informs that the limit was up to 14 days after the onset of symptoms and that the median was seven days .
Rômulo sent messages to the agency’s e-mail in April 2021, warning about the presence of these clear factual errors and a host of other inaccuracies and inconsistencies. He also made suggestions related to the editorial line, such as omission of facts and lack of clarity. The corrections and criticisms were ignored by the agency. After waiting about two months since the first contact, Rômulo decided to contact the IFCN, from the Poynter Institute, but he also never received a response. So he decided on his own to seek the report of Gazeta do Povo to tell his story.
“When seeing the fragility of the process, the poor quality of the checking service and the refusal to correct, probably motivated by political issues, I thought it was vital to bring this information to the surface”, says Rômulo about what his motivations were when seeking the press.
The report by Gazeta do Povo also received no answers. When asked about the criteria for selecting the content to be checked by Estadão Vera, about its funding sources and about the International Fact Checking Network (IFCN) process, Daniel Bramatti limited himself to replying: “Hi, Rodrigo! let’s speak out”, mistaking the name of the reporter who contacted us.
The agencies Lupa and Aos Fatos did not respond to our contacts.
Baybars Örsek , director of the IFCN, told Gazeta do Povo that the evaluation and review process takes place every year for each certified signatory organization and that it has already occurred for organizations to be kicked out of the network due to violations of its code of principles. Baybars also said that complaints submitted to the IFCN are reviewed by evaluators each year during renewals and that complaints are not addressed individually unless they show a pattern of violation of the Code. of Principles of the IFCN.
We request more information from Bayb ars about the organizations expelled from its network and about the lack of response to Rômulo’s question, but we did not get any further response.