The Effect of Social Media Platforms on Neglected Countries and Diaspora Communities: An Investigation into the Eritrean Case

RESEARCH QUESTION 

The research aims to provide a comprehensive picture of the manner in which social media platforms are used by Eritreans in the country and diaspora communities, including an analysis of the role of social media platforms in the Tigray conflict. The main research question is: What is the impact of social media platforms in countries and diaspora communities that are not considered important by the companies who own these platforms? The specific questions are: (1) How many accounts (on Twitter, Facebook, YouTube, Instagram, TikTok, Clubhouse, Telegram) are there that engage in any discourse related to Eritrea? (2) It has been reported that hate speech and misinformation in English, Spanish, French, and other Western-centric languages are much more likely to be taken down, while those in other underserved languages are not. What are the languages used, and from where? (3) How many accounts have proliferated hate speech, fake news, misinformation and disinformation, and what do they look like? (4) How well do automated tools perform in attempting to detect these issues? 

WHY IS THIS RESEARCH IMPORTANT? 

The research findings arm policymakers, citizens and impacted communities with information and recommendations to help regulate social media companies of mis/disinformation and ensure adequate resources are allocated to make products safe for all communities. This information is needed in order to prevent these platforms from being used to exacerbate tensions, particularly during conflict situations. The Internet was created with the hopes of connecting people, and if we are to have an Internet that promotes democracy and freedom for all, we have to have a deep understanding of how it is currently falling short. 

METHODOLOGY  

This research is a comprehensive quantitative and qualitative study to undertake a field-level inquiry into the social media landscape.  

In addition to collecting basic statistics on the presence/absence of particular keywords, phrases, and hashtags, the research relies on natural language processing (NLP) and social network analysis (SNA) strategies in the quantitative analysis. The analysis is conducted across multiple social media platforms, including Facebook, Instagram, Twitter and YouTube. This quantitative analysis is supplemented by qualitative interviews with journalists, activists, and victims of dis/misinformation, who include Eritreans, ethnic Tigrayans, and other people familiar with the regional context. 

The research involves collaboration with linguists to develop language technology for languages like Tigrinya, for which tools do not exist to perform any type of analysis. NLP classifiers shall be developed for identifying hate speech in relevant languages such as Tigrinya, which typically do not have a robust set of NLP tools. 

MEDIA AND PUBLICATIONS 

See below for research publications and other articles related to the research. 

https://www.fastcompany.com/90952740/why-timnit-gebru-wants-ai-giants-to-think-small

https://pml4dc.github.io/iclr2023/pdf/PML4DC_ICLR2023_39.pdf

https://www.techpolicy.press/policy-is-urgently-necessary-to-enable-social-media-research/