CoronaCheck and Fake News

With the spread of the Covid-19 virus, there has been a spike in disinformation about its origin, diffusion and effects. False and misleading claims concerning the 2019-20 coronavirus pandemic are documented in the Wikipedia entry “Misinformation related to the 2019-20 coronavirus pandemic.” People have died from taking drugs passed off as useful. The information verification process has been in trouble for some time, and the pandemic has made the problem plain for all to see.

Infodemia and data verification

The causes of this crisis have been known for a long time: people spreading incorrect information for political reasons, together with the great speed of news propagation on digital platforms. Given the large number of incorrect statements on the internet, it is very difficult to limit their circulation, because fact-checking takes considerable time and effort. To implement such checking it is necessary to identify a fact that is not simply an opinion, collect relevant data from reliable sources, and carry out the analysis to validate or otherwise what is claimed.

Unfortunately, given the overload of information on the internet, we are now facing what the World Health Organization (WHO) has called an “infodemia,” namely, the circulation of too much information, which makes it difficult to limit the spread of fake news. There is a lot of evidence that systems designed to moderate content distributed online are in great difficulty because of this problem.

Moreover, citizens are overwhelmed by false information, even that coming from people they trust privately – thanks to word of mouth communication and in our experience WhatsApp – and they do not have the tools to verify its correctness and reliability. This problem has surged with the coronavirus crisis, but it is a broader issue that applies to any topic beyond the health emergency, whether it involves industrial, economic or commercial issues.

In response to the information crisis, especially online, there are many initiatives that revolve mainly around the world of journalism. Especially in the English-speaking world, journalists have always considered fact-checking to be the mark of reputable editorial staff. As the number of alleged facts to be checked increases, organizations have been created where teams of fact-checkers verify facts (for example, snopes.com and politifact.com or, in Italy, pagellapolitica.it). These sites are independent of major newspapers, but there are also very effective initiatives within major newspapers, as in the case of lemonde.fr/verification.

The checks made by these experts are such a valuable resource that Google considers them among the most relevant results in related searches and Facebook buys their checks as a service to use them within its social network in order to identify false content. Obviously these are to be counted alongside a veritable army of thousands of human “moderators.” Recently, the problem has become so noticeable in Italy that the state broadcaster, Rai, and the government have activated a task force against misinformation.

The educational problem and research

The controversial reaction of many people in Italy to this news makes it clear that the problem is not only the scale, that is the large number of false “facts” and the impossibility of verifying them all manually. The great Italian controversies are in fact centered on the issue of the right to one’s opinion and the fear that someone from above can decide what is true and what is false. This is also an educational problem because many people do not know what objective verification is. It is said that everyone has opinions, and that all opinions count in the same way. Whenever we talk about objectivity, some will think of methods of control that violate the right to expression. If, on the one hand, the crisis is educational, on the other hand there are initiatives to resolve it from different angles. Certainly this is the case with education itself (see, for example, factcheckingday.com) and with the creation of accessible resources and content, but also with a computational approach to the problem.

For years the scientific community has focused on the problem with conferences and journals dedicated to combating disinformation and propaganda, especially online. Many approaches have been proposed to automatically verify different types of facts, to track and understand how to slow down the proliferation of false news on social networks, to limit the effects of so-called “bots” (short for robots), i.e. programs that automatically re-launch content, and many other aspects.

The CoronaCheck

These efforts are particularly relevant and important because they can help to combat both of the problems we have raised. We want to signal one that, in the case of the coronavirus, is proving to be very significant. It comes from the work of two researchers: Prof. Paolo Papotti from Eurecom University (France) and Prof. Immanuel Trummer from Cornell University (USA). They have developed a computer system that automatically verifies claims about the coronavirus. CoronaCheck (also available in English at https://coronacheck.eurecom.fr/en) is a website where facts can be verified through official data. For example, given a claim like “Mortality in Italy is much higher than in France,” the system determines whether this is true or false. Each statement is verified on the basis of official data collected daily, thanks to Johns Hopkins University, from sources such as the World Health Organization, governments and health ministries of the various nations.

In addition – and this is really relevant – the system learns from user feedback how to manage new types of statements and how to exploit new data sources. The system recognizes statements that it cannot verify and, in these cases, asks the user to help in the process. This interaction creates new examples from which the model learns to verify new types of statements over time.

The CoronaCheck is a very helpful tool and can be used by large networks to identify and limit false claims before they become popular, and also by any citizen who wants a reliable tool to verify the information he or she receives.