The AI-powered tool monitors the spread of hate speech and violence in both Hebrew and Arabic within the Palestinian context.
This statement was originally published on 7amleh.org on 15 July 2024.
7amleh – The Arab Center for the Advancement of Social Media, launched the “Arabic Violence Indicator”, a linguistic model that uses artificial intelligence technology to automatically and immediately monitor the spread of hate speech and violence on social media platforms in Arabic Language. The findings are then presented on “7or – the Palestinian Digital Rights Violations Observatory”; this is a significant step to fight violent content and protect users from online threats. This new tool follows the successful release of the Violence Indicator for Hebrew content in November 2023.
7amleh has developed AI-powered Language Model classifiers to monitor the spread of hate speech and violence in both Hebrew and Arabic within the Palestinian context. The new Arabic Violence Indicator provides real-time statistical data on the reach, distribution, and engagement with harmful content.
The Arabic Violence Indicator detects violent content and hate speech on social media platforms among Palestinians, evaluating factors such as gender, political background, religious background, and more. This tool not only identifies harmful content but also analyzes its distribution and engagement, helping us understand and mitigate its impact.
The Violence Indicator helps researchers, decision-makers, human rights defenders (HRDs), activists, media professionals, and organizations understand the extent of online threats and violence. It enables them to study these phenomena in-depth and design effective interventions to combat such harmful content.
“The Arabic version of the Violence Indicator represents a significant leap in 7amleh’s efforts to combat violent content and online threats within Palestinian society. This tool will help us understand the prevalence of such content and provide a clearer picture of its potential complications and threats, especially for active online users and the connection to offline harmful attacks.” Ahmad Qadi, Monitoring and Documentation Manager at 7amleh said. “This strategic step will enhance our advocacy for better policies, protect users, and provide valuable data for stakeholders to design more effective interventions to safeguard people online.” He added.
Since January 2024, the Arabic Violence Indicator has detected around 261,000 pieces of violent content in Arabic retroactively. The majority of this content was found on Facebook (88.6%), with X (formerly Twitter) accounting for 5% and Telegram for 6.4%. Notably, 48.6% of the detected violent content was based on gender.
