The Active Role of Platforms During the Pandemic: A Review of Their Policies and the Gaps That Are Beginning to Emerge

The outbreak of the COVID-19 pandemic has left the entire world on edge. While researchers, doctors, and scientists are focusing on possible solutions or measures to slow the spread of the virus,

The outbreak of the COVID-19 pandemic has left the entire world on edge. While researchers, doctors, and scientists are focusing on possible solutions or responses to slow the spread of the virus, the rest of the public is also grappling with the issue in one way or another. People around the world have seen their routines disrupted and have had to adapt to work dynamics that differ from the norm. Many have seen their work responsibilities, family and social relationships, and leisure time shift to the digital realm. Against a backdrop of significant online presence and engagement, as well as widespread use of mobile devices, the circulation of content on social media has also been impacted by the emergence of COVID-19. What role have the major platforms played in recent weeks? Generally speaking, it could be said that they have maintained a strong presence and played an active role in moderating the content circulating on their platforms. Some even argue that they are taking control of the information ecosystem as never before. Nevertheless, major companies have sent their employees to work from home, leaving a greater flow of content in the hands of artificial intelligence and machine learning tools. Let’s look, for example, at some of the measures Facebook and Twitter took during the pandemic. In March, Facebook began promoting its “COVID-19 Information Center” (COVID-19 Information Center) on its platform, located at the top of the news feed and featuring real-time updates on measures taken by national authorities and global organizations such as the World Health Organization (WHO). In the press release, the company also noted that it was “carefully monitoring user patterns,” “making systems more efficient,” and “adding capabilities as needed.” However, in their effort to prioritize and provide reliable information, it is not entirely clear how they will respond to government measures that do not follow the recommendations of the WHO or of sources they consider authoritative or credible on the matter. This would be the case, for example, with the Nicaraguan government. As for misinformation, Facebook has been actively working to remove any COVID-19-related content that could “contribute to imminent physical harm.” In this regard, another statement issued in late March indicates that since January, they have removed posts that “make false claims about cures, treatments, the availability of essential services, or the location and severity of the pandemic.” An example of a post that could be removed for this reason is one that claims that physical distancing does not help prevent the spread of the virus. Furthermore, they have also begun working with fact-checkers to address what they identify and refer to as content that does not directly result in physical harm but may cause social unrest, such as conspiracy theories about the origin of the virus. In accordance with this policy, a few days ago the company removed a post by Brazilian President Jair Bolsonaro, which sparked controversy and mixed reactions. In the video, the president claimed that “hydroxychloroquine works everywhere,” despite warnings from health authorities that the drug is still undergoing testing to determine its effectiveness. Facebook determined that this content, regardless of who posted it, violated the social network’s rules—or Community Standards—and therefore removed it. Similarly, Twitter has issued public statements regarding the adaptation its policies in response to the outbreak of the pandemic. The company stated that it has broadened its definition of “harm” to include content that directly contradicts the recommendations of global and local health authorities. Although Twitter takes an extremely open stance toward content of public interest (that is, tweets remain even if they violate the social network’s rules if they are valuable to the “health of public conversation”), it did not show such tolerance toward content posted by Bolsonaro and Venezuelan President Nicolás Maduro. Both leaders made statements that contradicted WHO guidance, so their tweets were removed. Given this situation, questions arise about the gaps that still exist in these platforms’ policies. A few days ago, Agustina Del Campo commented on this blog that there is a debate—predating COVID-19—regarding how companies define and classify the discourse of public officials. The examples cited above clearly highlight the importance of this discussion. But they also raise new questions: What content is being removed? Are there—or will there be—reports that can shed light on content moderation and its potential effects during the pandemic? If, as Facebook announced, its content moderators are working remotely and, as a result, the use of artificial intelligence and machine learning has increased to flag and remove content that violates the platform’s policies, to what extent is it possible to determine the legitimacy of the content? There has been an ongoing debate on this issue, predating the pandemic, regarding the use and appropriateness of automation to resolve content issues that require analysis of context, local idioms, current events, and so on. Therefore, it should come as no surprise that there are currently voices (for example, Emma Llanso of the Center for Democracy and Technology) calling on companies to produce a detailed analysis, once the crisis ends, of how their systems performed amid forced automation. While some companies (Twitter, Facebook, YouTube) made it clear in advance that much of the content would be disproportionately affected even despite their best efforts, it would be essential to know what types of content were most prone to errors or which voices were silenced or restricted by automation during this period. The answers to these questions will help companies objectively evaluate the performance of their algorithms and systematize some of the data for which we currently have only anecdotal information. And they could, in turn, serve as an empirical basis for the proposed regulation of platforms that has been so widely discussed globally over the past two years. By Paula Roko TW @paulaaroko