A multilingual analysis of pro Russian misinformation on Twitter during the Russian invasion of Ukraine.

The Russian government has long since engaged in an information campaign of propaganda and disinformation as a major part of foreign policy. This has been taken to new heights since the invasion of Ukraine in February 2022. In this study, we investigate pro-Russian misinformation within the opening weeks of the invasion in 6 languages: English, Japanese, Spanish, French, German, and Korean. Using Twitter data, we apply a combination of network and language embedding models to identify popular topics of misinformation amongst users in each language. Despite English users forming the most dominant language base on Twitter, we find that the popularity of misinformation in Japanese regularly outstrips English for certain topics. Misinformation shared by Spanish users is also over-represented in proportion to its much smaller user base. Our results provide insight into the current state of misinformation in each language. While we discuss some of the possible drivers behind the factors such as language over-representation, our study also highlights the need for further cross-lingual misinformation research in order to better understand this phenomena in a truly global context.

A cross-lingual analysis on the spread of misinformation using the case of Ivermectin as a treatment for Covid-19.

The spread of misinformation transgresses international boundaries, between languages and cultures. This is especially evident in times of global crises such as the Covid-19 pandemic. This study observes misinformation on Twitter in the Japanese and English languages regarding false claims that the drug Ivermectin is an effective treatment for Covid-19. Our exploratory cross-lingual analysis identifies key themes of discussion and influential users in both languages, finding English misinformation to be highly popular amongst Japanese users. Significantly, an analysis of the timing of retweets between languages reveals that Japanese users find and widely share English misinformation often before English users themselves. This contradicts expectations that users from other languages tend to pick up on popular misinformation in English. Instead, they seek out English language sources irrespective of their popularity to support their agenda. These results emphasise the importance of cross-lingual mitigation strategies for organizations trying to combat misinformation, and that they must look beyond their own language spheres.