Governments’ use of AI systems in social media monitoring during elections

Based on a thorough analysis of reports from the Office for Democratic Institutions and Human Rights (ODIHR) on election monitoring efforts in EU member countries between 2016 and 2023, we have identified alarming trends on social media. These trends include the spread of hate speech, misinformation, and targeted attacks on women, which pose significant threats to the integrity of electoral processes and fair political discourse.

Key Findings:

1. Rising Hate Speech and Misinformation: Our analysis reveals a troubling increase in hate speech and misinformation spread through social media platforms during electoral periods.
2.  Harsher Discourse Compared to Traditional Media: Social media discourse tends to be harsher and more polarised than traditional media channels. 
3.  Targeting Women with Negative Messaging: Women candidates and politicians are disproportionately subjected to negative messaging on social media platforms.

This environment may discourage meaningful political participation and compromise the quality of public debate. To effectively address these challenges, the following policy recommendations are proposed:

Policy Recommendations:

1. Monitoring social media: OSCE ODIHR should establish comprehensive monitoring mechanisms for detecting and mitigating hate speech and misinformation on social media platforms.
2.  Utilisation of AI for Monitoring: OSCE ODIHR should use AI-powered tools to monitor social media platforms. AI can enable more proactive and accurate detection of harmful content, allowing for more timely detection of emerging threats.
3.  Gender-sensitive Approach: OSCE ODIHR should develop gender-sensitive strategies to address the targeted attacks on women in social media discourse.