Russia leading from behind: Coronavirus-focused case study of cross-border disinformation spread

Disclaimer: This report is an independent product of Semantic Visions, shared with the Global Engagement Center (GEC) at the United States Department of State to enhance international understanding and responses. As the work of Semantic Visions, this report does not necessarily reflect the views or analysis of the GEC or the Department of State.

This report is available as PDF


Semantic Visions (SV) has conducted an in-depth assessment of cross-border transmission of disinformation related to the global outbreak of coronavirus. Using the company’s proprietary military-grade Open Source Intelligence (OSINT) System, the report looks at the situation from a global perspective using the case study of the Czech Republic as a target country. It analyses the origin, development and shaping of disinformation narratives and tracks their pathways into the Czech-speaking cyberspace. The information used for this report is based on advanced semantic and sentiment analysis system that processes over 1 million news articles per day, from 750 thousand online sources.

Initial reporting on the virus

The Chinese officials announced the outbreak of a new virus in Hubei Province on the 7th of January 2020. However, the Semantic Visions OSINT system was able to identify an existence of a yet unknown pneumonia-related virus before that through its threat tracking of “Epidemic” and “Infections” related risks.

The very first piece of news on the outbreak surged on the 31st of December 2019 and was based directly on the report of the Wuhan Municipal Health Commission.

The above chart shows the beginning of media coverage on the existence of a yet unknown pneumonia related virus from Chinese (red) and English (blue) sources on the 31st of December 2019.

We see that Chinese-speaking media were first to report on the outbreak. While Chinese-speaking sources outside of People’s Republic of China were the first ones to react, mainstream media in Mainland China reported on the topic only hours later.

Using trend/pattern comparison of the first several weeks of news coverage on the new coronavirus disease published in English and Chinese, we see very similar trends/patterns. This points to a surprising level of openness of Chinese media at this stage, especially when compared to the reporting on the outbreak of SARS in 2002.

The above charts show the growth trend/pattern of early-stage media coverage in Chinese.

The above charts show the growth trend/pattern of early-stage media coverage in English.

First disinformation – Russia and the U.S.

The first disinformation narratives quickly followed the official announcement of the virus. According to research conducted by Semantic Visions, the first virus-focused disinformation began appearing in the Russian-speaking cyberspace as early as 20th of January 2020 (firstly published by, a Russian state-owned media operated by the Russian Ministry of Defense), accusing the US of producing COVID-19 as a biological weapon. Before this article, Semantic Visions has not identified any similar conspiracy about COVID-19 being a man-made weapon.

Another surge of disinformation happened just shortly after on the other end of the planet, this time originating in the United States alt-right web news scene. Unlike the Russian sources that suggested that the virus was made by the US military, the American websites implied that the virus was accidentally leaked from a Chinese top-secret governmental biological warfare lab in Wuhan.

The narrative of the virus as being an artificial biological weapon, designed and deployed against China by the US government, was later taken up and amplified by the Chinese governmental propaganda.

The story became an important tool used to divert public attention away from accusations of the initial mishandling of the crisis by the Chinese government. This became particularly noticeable when the narrative was picked up by Chinese governmental officials. The fact that Chinese propaganda amplified a narrative created by the Russian disinformation ecosystem indicates, at the minimum, a convergence of Russian and Chinese interests in the information system.

The above dashboard shows the growing trend of the biological weapon narrative in the world.

Disinformation spill-over to the Czech republic

Disinformation-oriented media in the Czech Republic were a bit slow to pick up the topic at the start. The first mentions of the virus were linked to the Gates Foundation Event 201 simulation and New World Order conspiracies, which imply that global elites knew about the virus outbreak months before it began. The first pieces of news that clearly talked about the artificial origin of the virus appeared in the first week of February. Interestingly enough, the first piece deployed both narratives – Chinese lab leak as well as a U.S. biological weapon attack against China.

The origin of both narratives used in the Czech disinformation environment can be quite clearly tracked, as both the Russian media and the US alt-right scene quoted different “experts” on the topic. In Russia, nearly all coronavirus disinformation, including the very first disinformation on this topic mentioned above, quoted so-called “expert” Igor Nikulin. In the US, it was Francis Boyle’s claims on coronavirus being a Chinese biological weapon made on 30th of January, that quickly spread in the alt-right scene.

Czech disinformation actors (who have without exceptions pro-Russian views) began citing content from both the US and Russia, including quotes of both of those personas as soon as in the second week of February. This implies that they used content from both Russian pro-Kremlin outlets and US alt-right media.

Overview of narratives

Although the narrative of a virus as an American-made biological weapon became a dominant one, the entire scope of narratives present in the Czech discourse turned out to be more diverse.  The narratives in the Czech cyberspace included claims that the spread of the virus in the EU is linked to migration and to US soldiers moving around the continent as part of the Exercise ‘Defender Europe 20’, a US Army led, joint, multinational training exercise including NATO’s participation. The EU was accused of purposefully ignoring the coronavirus crisis.

Thus, the messaging about the virus copied the usual bias of the Czech disinformation-oriented outlets: anti-EU, anti-US and pro-Russian messaging alongside with the stoking of fear for the alleged mass migration threat.

The disinformation outlets also accused the Czech public media of purposefully exaggerating the virus threat in order to attack China and divert attention from another soon-to-happen migration crisis. This fits into the specific context of Czech domestic politics, where public service media, specifically the Czech Television and the Czech Radio, are under constant pressure from pro-Russian and pro-Chinese domestic actors.

These actors, including the disinformation media, but also the incumbent President of the Czech Republic Miloš Zeman, are currently using the crisis to further enhance Chinese influence and undermine its opponents in the country. This is further amplified by the news that began to appear in the Czech disinformation media in March, portraying China as a humanitarian superpower saving the European countries struck by the virus.

What does the development of coronavirus disinformation in the Czech context tell us about the spread of disinformation across borders and languages spheres? What we can establish with certainty is that Russian media picked up on the topic of COVID-19 as an artificial weapon before anyone else did, and quickly established the initial anti-American, man-made virus narrative. This narrative was from the beginning and in almost all cases accompanied with citing of the so-called “expert” Igor Nikulin.

Czech-speaking disinformation websites mention the coronavirus only two weeks after the topic appears in the Russian outlets and they began referring to Nikulin’s claims three weeks later than their Russian counterparts. Firstly, the Czech-language sources cited both disinformation stories – the one that originated in Russia on the 20th of January, and the one that originated on the US 26 of January. However, what prevailed in subsequent reporting was the disinformation with the anti-US angle.

To make things a bit more complicated, the Russian sources took over also the disinformation story that originated in the US only days after its publishing. This is in line with the EEAS report findings on Russian disinformation about the coronavirus which mention a change of tactics by pro-Kremlin media, switching from authoring disinformation narratives to amplifying theories that originated elsewhere – to create a digital smokescreen that makes it more difficult to track the origin of the disinformation, and to give the pro-Kremlin media the option of plausible deniability.

In our opinion, mixing the Molotov disinformation cocktail with often contradictory messages corresponds with the Russian disinformation strategy, which is to confuse people and undermine the public trust of governments, thus imposing chaos in Western democracies.


The key takeaways from the research conducted by Semantic Visions are following:

  • Trend/pattern comparison of English and Chinese reporting reveal surprising level of openness of Chinese media at the early stage of outbreak. Particularly in comparison with the Chinese media reporting during the outbreak of SARS epidemic in 2002.
  • Russian sources were first among the global disinformation to pick up on the new topic, with U.S. sources following several days later and Czech sources following weeks later.
  • The narrative of the virus as an artificially created biological weapon designed by the U.S. military and deployed against China was originally authored by Russian-speaking sources.
  • Czech disinformation news outlets took over content and narratives from Russian sources. However, there was a weeks-long delay in-between and the Czech outlets also used content from U.S. based sources. This would suggest that the connection between disinformation outlets in Russia and in Central European countries might be less direct than it was previously assumed.
  • The assumption made by the EEAS, the EU body in charge of countering disinformation, is not supported by our research. The case of Russian media re-circulating the Washington Times article from the 26th of January would support the EEAS findings. However, the fact that the Russian outlets were the first ones to create and distribute the “virus made in the U.S.” narrative would suggest otherwise.
  • The most likely conclusion is that both of the approaches are now applied simultaneously: while Russian propaganda still continues to author its own narratives, it might increasingly amplify those originating elsewhere.
  • Such a dual approach to spreading disinformation corresponds with the Russian strategy of waging information warfare. By creating a figurative Molotov “disinformation cocktail” through mixing multiple, contradictory messages, Russia seeks to confuse people, undermine trust in their governments and thus impose chaos on Western democracies.

About Semantic Visions

Semantic Visions (SV) is a data analytics and risk assessment firm based in Prague, Czech Republic. SV runs a proprietary military-grade Open Source Intelligence (OSINT) System. SV’S team has over 15 years’ experience in OSINT data collection and automatic understanding of textual information across the world’s top languages. SV is one of the very few organizations in the world able to see the big picture of the global online news sphere and has long-term practical experience in detecting hostile propaganda and disinformation. In March 2019, SV took first place in the US-UK Tech Challenge for its development of proprietary technology that can identify and track disinformation and propaganda around the globe.

SV risk solutions have been successfully integrated into the world’s largest business commerce network SAP Ariba, which drives over US$3 Trillion in commerce – more than Amazon, eBay and Alibaba combined. SV delivers real-time risk detection covering millions of companies and thousands of geolocations.