The Pandemic and Combating Misinformation

COVID-19 also sparked a pandemic of misinformation on social media. Conspiracy theories, miracle cures, alternative or experimental treatments, and healing blessings are the order of the day. L

COVID-19 also sparked a pandemic of misinformation on social media. Conspiracy theories, miracle cures, alternative or experimental treatments, and healing blessings are the order of the day. The lines between information and opinion are blurring, and social media platforms are taking emergency measures that, however, raise many questions regarding freedom of expression. In March, Twitter, Facebook, and YouTube removed videos posted by Brazilian President Jair Bolsonaro in which he urged people to break quarantine and spoke of hydroxychloroquine (a drug used to treat malaria) as an effective remedy for the virus. Twitter also removed a tweet by Rudy Giuliani, Donald Trump’s personal attorney, in which he claimed that this treatment was “100 percent effective”; it removed a tweet by Nicolás Maduro in which he proposed a home remedy for the virus; and temporarily suspended the account of the conservative magazine The Federalist, after deleting a tweet from him that linked to a post suggesting treating the pandemic by intentionally infecting people. Are these suspensions censorship, an attempt to prevent the spread of the coronavirus, or a bit of both? For years, these platforms have faced constant demands to monitor the content posted by their users. However, this demand raises two issues: (i) a regulatory one: platforms are tasked with defining the limits of free speech and (today) determining what aligns with—and what does not align with—the current scientific paradigm; and (ii) an implementation challenge: the effective enforcement of platform rules to monitor the millions of posts made daily on the internet is impossible. Aware of the immense (and costly) technical and human effort involved in implementing these rules, and reluctant to assume the role of arbiters of truth, platforms have been lax and cautious when it comes to controlling misinformation. For example, Twitter had no specific rules on this matter on its platform—focusing on account activity rather than content—and Facebook had relied primarily on the work of external fact-checkers… until now. Twitter had to rush to adjust its rules to address the situation, shifting from having no policy on disinformation to an aggressive strategy that is stricter than Facebook’s. Now, the platform prohibits content that clearly aims to incite action that could pose a direct risk to people’s health and that contradicts the guidance provided by “authoritative sources of global and local public health information.” But the rule is so ambiguous that it doesn’t clarify who might be considered an authorized source of information or what happens when there are contradictions among them (Bolsonaro vs. the WHO or Bolsonaro vs. his own health minister, for example). Among other things, Twitter no longer allows users to deny the recommendations of health authorities (“social distancing isn’t effective”); describe treatments as “ineffective,” even if done as a joke or if they’re harmless (on Facebook, these are allowed, but warnings are added and their visibility is reduced) or denying scientific data on transmission during the incubation period. In other measures, Facebook has decided to ban all advertisements for the sale of face masks, hand sanitizer, disinfectant wipes, and coronavirus test kits. For its part, in a drastic move, Google is not allowing any ads that include words related to the virus, even if they are beneficial to users or provide information of public interest, such as ads for donation channels. The platforms are acting as if there were a consensus on key issues regarding the spread of the virus. For example, while people were calling for a ban on international flights, the WHO advised against closing borders. At the start of the pandemic, we were told that only infected people should wear masks; today, we’re told that everyone should. Experts also don’t seem to agree on whether the virus spreads through the air or not, not to mention the effectiveness of homemade face masks, recommended by the U.S. Centers for Disease Control and Prevention. To make matters more difficult, thousands of people who work on enforcing social media policies (that is, reviewing and removing potentially infringing content) are also working from home to comply with social distancing guidelines. As a result, platforms have to rely more and more on their artificial intelligence tools to make these kinds of decisions. The challenge is immense: they must remove prohibited content without harming information of public interest, relying on algorithms that do not understand context and that must be adjusted in record time. Frustrations and errors were not long in coming. In March, Facebook blocked hundreds of legitimate posts related to the coronavirus, after they were mistakenly flagged as spam. More recently, Facebook’s system threatened to block people who are organizing groups to make homemade face masks, but the social network later clarified that it was all due to an error in its artificial intelligence technology. Furthermore, the social network has not only asked users for patience regarding errors and delays but has also stated that it will likely not review users’ appeals of its moderation decisions. In the classic debate over online content moderation, there’s always the question of whether we really want to entrust platforms with determining what is acceptable and what isn’t. However, the changes that have taken place in recent days raise new questions. It’s common for governments to implement regulations during emergencies that are later made permanent. These days, there is much talk about surveillance technologies that will remain in place and continue to operate after the emergency. Will the same be true of the new rules governing speech after the pandemic? Likewise, now that artificial intelligence technologies have been developed to address the crisis, to what extent will we continue to rely on them rather than on human oversight, which can understand context? It seems as though our words are destined to be controlled by machines. Finally, the measures taken by platforms are affecting legitimate forms of expression—to what extent will freedom of expression be sacrificed? By Luisa Fernanda Isaza. Luisa is a lawyer and researcher at Linterna Verde and a consultant for the Legislative Observatory in Colombia. This text is part of a project on content moderation between Linterna Verde and CELE. Twitter @luisaza * This text was originally published in La Silla Llena, a discussion forum hosted by La Silla Vacía.