A Data-driven Understanding of Left-Wing Extremists on Social Media
More languages
More actions
"A Data-driven Understanding of Left-Wing Extremists on Social Media" is a July 2023 paper published by Utkucan Balcı and Jeremy Blackburn of Binghamton University and Michael Sirvianos of the Cyprus University of Technology. It focuses on the "tankies," a term which it uses 399 times over the course of 38 pages.[1]
Content[edit | edit source]
Horseshoe theory claims[edit | edit source]
The paper says that both tankies and right-wing extremists migrated to new platforms after being banned from Reddit and increased their use of "toxic language." It also notes that both leftists and rightists fought against ISIS in Syria.[1]:2–3
Misunderstanding of Marxist positions[edit | edit source]
The paper incoherently describes tankies as "more Stalinist than Leninist."[1]:1
Based on AI analysis, the article falsely states that tankies claim all Jews support Zionism and that tankies attack Jews for their identity.[1]:18 It also claims that tankies attack the identities of Africans, Arabs, Asians, Hindus, Mexicans, and whites in over 20% of posts mentioning these identities,[1]:21 which combined would mean that they attack most people in the world for their ethnicity.
Western terminology[edit | edit source]
The paper abbreviates the Communist Party of China the CCP instead of CPC[1]:17 and calls Russia and China "revisionist world powers."[1]:22
Methodology[edit | edit source]
When finding the most discussed topics, the paper does not filter out basic grammatical words such as 'the,' 'is,' and 'of.'[1]:13
The paper tries to use the Perspective API to numerically measure how the toxicity of Reddit comments in order to graph them.[1]:19–20