Shifting Patterns of Extremist Discourse on Facebook: Analyzing Trends and Developments During the Israel-Hamas Conflict
Researchers analyzed 12 months of hate group engagement across 5 major war events.
A new arXiv preprint (2605.01528) by Rr. Nefriana and seven co-authors from institutions including the University of Pittsburgh analyzes shifting patterns of extremist discourse on Facebook during the Israel-Hamas conflict. The study examines 12 months of data (July 2023 to June 2024) from Facebook groups categorized as anti-Israel/Semitic, anti-Palestine/Muslim, and anti-both. Engagement, sentiment, and topic trends were mapped against five major conflict events, including the October 7 attack and Israel’s withdrawal from Khan Yunis in Gaza. The findings confirm that shifts in hate speech activity correspond closely with real-world events, but the patterns differ significantly across group types. At the conflict’s onset, the proportion of activity in anti-both groups fell, while single-sided hate groups (anti-Israel and anti-Palestine) increased their share. This trend reversed after the Khan Yunis withdrawal, suggesting that binary opposition drives more engagement during acute phases. Across all categories, negative content proportion surged, and neutral content dropped sharply. Topic modeling revealed that anti-Palestine/Muslim groups pivoted from religious discourse toward social media activism and political protest language around the war’s start, while anti-Israel/Semitic groups shifted from political/protest to religious topics two weeks before the war. The paper provides quantitative evidence that extremist online ecosystems react asymmetrically to conflict events, with implications for content moderation and counter-speech strategies. The authors note these patterns could inform automated detection systems and policy interventions tailored to specific hate group behaviors.
- Anti-both hate groups saw a decrease in activity proportion at the conflict's onset, while one-sided hate groups increased; the pattern reversed after Israel's troop withdrawal from Khan Yunis.
- Negative content proportion surged and neutral content fell across all three group categories during the conflict.
- Anti-Palestine/Muslim groups shifted from religious to social media activism and political/protest topics as war began; anti-Israel/Semitic groups moved from political/protest to religious topics two weeks before the war.
Why It Matters
Real-time hate speech monitoring systems can use these event-driven pattern shifts to improve detection and moderation.