The contours of state-aligned gendered disinformation online

Abstract

Online spaces are being systematically weaponised to exclude women leaders and to undermine the role of women in public life. Attacks on women which use hateful language, rumour and gendered stereotypes combine personal attacks with political motivations, making online spaces dangerous places for women to speak out. And left unchecked, this phenomenon of gendered disinformation, spread by state and non-state actors, poses a serious threat to women’s equal political participation. In this research, we investigated state-aligned gendered disinformation in two countries, Poland and the Philippines, through an analysis of Twitter data. We analysed tweets in Polish and, from the Philippines, in English. All Twitter data was collected through the platform’s public API. Data was collected and analysed as follows: 1. A list of users who heavily influence online political discourse in each country was assembled, and tweets sent by these accounts and a sample of users following them was collected. 2. Tweets were filtered to those containing terms within a lexicon of political and gendered terminology, which was built by in-country experts and refined according to language discovered in the dataset. 3. The resulting dataset was filtered to identify tweets containing a name or phrase associated with a likely target of gendered disinformation, based on initial analysis and input from incountry experts. 4. This data was manually analysed to identify which tweets displayed gendered disinformation, and the key themes which arose. 5. Finally, the dataset was used to generate network maps of which users were sharing gendered disinformation, and whether there was evidence of state-alignment in those who were most engaged in this activity. In August 2020, a closed roundtable was held by NDI with experts from across the world to discuss the findings and the ramifications for mitigation strategies. What did we find? • Gendered disinformation is being shared by state-aligned actors online - though it reaches much broader audiences. • The way in which gendered disinformation spreads across a network varies hugely according to context. • But the themes of gendered disinformation - the rules that it follows - are often broadly consistent. • Gendered disinformation is parasitic on news events, existing rumours, and underlying social stereotypes, and seeks to reshape the terms of political discourse in a way that harms women. • Gendered disinformation plays on existing tropes to try to convince people that women in public life, are one or more of: devious, stupid, overly sexual, in need of protection, or immoral: and so unfit for public life. EXECUTIVE SUMMARY ‘Attacks on women which use hateful language, rumour and gendered stereotypes combine personal attacks with political motivations, making online spaces dangerous places for women to speak out.’ 6 • Gendered disinformation is not just false information - it uses highly emotive and valueladen content to try to undermine its targets. • Gendered disinformation weaponises harassment against women in public life, and tries to make them afraid to talk back. • But women are speaking up: online counterspeech is being used by women in public life and their allies to fight back. What does this mean for the fight against disinformation? • Gendered disinformation should not be overlooked in responses to general disinformation. Solutions in policy and practice must recognise that disinformation takes many forms, and can vary what it looks like and where it originates according to context. • Centralised automated solutions to gendered disinformation are likely to censor legitimate speech and overlook gendered disinformation. Input and oversight from local experts who understand the language and context in which disinformation occurs is vital. • Solutions should centre and learn from the experiences of women who are already working to challenge disinformation: the problem is a systemic one and targets of gendered disinformation should not be expected to fix it as individuals.