Skip to content
Submit incident
AII-1000-58663
Well documentedThree or more independent publishers report the core facts.

Meta's content moderation delays linked to incitement in Bangladesh

Amnesty International warned that Meta's delayed content moderation on Facebook failed to prevent the spread of inflammatory and sectarian content in Bangladesh ahead of elections, contributing to a volatile environment that fueled mob attacks on media outlets and risked violence against minority communities.

18 December 2025 Β· Meta Platforms Β· Harmful and abusive content
18 December 2025
AII-1000-586633 sources ↓
Quick facts
Category
Harmful and abusive content
Country
Bangladesh
Source
Amnesty International
Date
18 December 2025
Drafted by
Raidu Incident Detection

What happened

Ahead of Bangladesh's 12 February 2026 parliamentary elections, Amnesty International observed a rise in harmful online content on Facebook, including misleading and inflammatory narratives targeting political parties and minority communities, and sectarian content designed to incite division. Much of this content originated from outside Bangladesh, particularly from India.

On 18 December 2025, violent mobs attacked the offices of The Daily Star and Prothom Alo, two leading Bangladeshi media outlets. Investigations by The Daily Star and Dismislab found that months of online threats and incitement on social media preceded the attacks, with the outlets being portrayed as 'Indian agents' and 'anti-national forces' and calls made to burn and attack their offices.

Bangladeshi authorities reportedly warned Meta about delays in addressing posts calling for violence and expressed concern about the impact on public security and minority communities.

Reported impact

Amnesty International warned that the combination of cross-border harmful content, political tension, sectarian narratives, and algorithmic amplification creates a volatile environment that could heighten the risk of sectarian tensions, discrimination and violence, particularly against minority communities and freedom of expression.

A direct link was found between online incitement of violence on social media and the mob attacks on the two media outlets.

Index analysis

The incident reflects a dangerous trajectory seen in multiple countries where online incitement, misinformation, disinformation, and coordinated harassment campaigns, when amplified by platform algorithms, can quickly spill offline into discrimination, violence and other human rights abuses.

Governance controls

Meta's content moderation system failed to timely address posts calling for violence on its Facebook platform, according to reports from Bangladeshi authorities and investigations by local organizations.

Classification

Category
Harmful and abusive content
Industry
Media, social platforms and marketing
Country
Bangladesh
Harm status
Actual harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Human or fundamental rights, Public interest
Harmed
General public, Business, Civil society
Failure modes
Content moderation failure
AI systems
Facebook content moderation system

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

Amnesty International
Meta content moderation shortcomings risk Bangladesh violence
16 March 2026
Prothomalo
Amnesty warns: Harmful facebook content raising risk of real-world violence in Bangladesh
17 March 2026
JURIST
Rights group warns Meta moderation policies fuel violence in Bangladesh
19 March 2026