Facebook allegedly did not adequately remove anti-Rohingya hate speech, some of which was extremely violent and dehumanizing, on its platform, contributing to the violence faced by Rohingya communities in Myanmar.
.avif){kind=small}
Facebook and WhatsApp were used as primary tools for spreading anti-Rohingya hate speech, dehumanizing propaganda, and fabricated news in Myanmar. This content, often orchestrated by the military and nationalist groups, incited real-world violence, including ethnic cleansing, displacement, and the assassination of prominent figures. Despite repeated warnings from researchers and NGOs, the platform failed to adequately police content or provide sufficient moderation resources in the local language, leading to significant human rights abuses.
Facebook allegedly did not adequately remove anti-Rohingya hate speech, some of which was extremely violent and dehumanizing, on its platform, contributing to the violence faced by Rohingya communities in Myanmar.
The Myanmar military systematically used Facebook to conduct a coordinated, large-scale disinformation and influence campaign targeting the Rohingya minority to manipulate public opinion and incite ethnic cleansing.
The risk was propagated and amplified by Facebook's automated content recommendation algorithms, which unintentionally facilitated real-world violence while pursuing user engagement goals.
Unacceptable Risk: The reports describe the platform's algorithms being exploited for large-scale behavioral manipulation and influence that directly contributed to real-world violence and ethnic cleansing.
Severity ratings are based on this scale.
Highest direct severity in any category: Severe. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.
Reported: The reports explicitly describe physical harm, including mass killings, rapes, and the assassination of prominent figures.
Indirectly caused: The military crackdown and communal violence, fueled by Facebook propaganda, led to the killing of dozens of people, mass rapes, and the assassination of prominent lawyer Ko Ni. In India, false WhatsApp messages led to the lynching and death of at least 7 people.
Inferred additional harm: It is highly likely that thousands of additional undocumented deaths and injuries occurred during the military's ethnic cleansing campaign against the Rohingya, which was actively promoted on the platform.
Reported: The reports describe villages burned to the ground and homes destroyed.
Indirectly caused: Buddhist mobs destroyed homes and forced villagers to flee in Bago, and military operations burned down entire Rohingya villages.
Inferred additional harm: Widespread destruction of community infrastructure, schools, and mosques in Rakhine State, though exact numbers are not fully quantified.
Reported: The reports describe the burning of villages and destruction of homes.
Indirectly caused: Mob violence and military crackdowns resulted in the burning and destruction of thousands of homes and properties belonging to Rohingya and other Muslims.
Inferred additional harm: The total loss of property for over 730,000 fleeing refugees is immense, likely totaling hundreds of millions of dollars in unquantified property damage.
Reported: The reports mention a 150 billion dollar class-action lawsuit filed by Rohingya refugees against Meta, but do not specify direct commercial financial losses of the victims.
Indirectly caused: Refugees lost their livelihoods, businesses, and assets when fleeing Myanmar, though specific average financial loss figures per occurrence are not detailed.
Inferred additional harm: The economic displacement of nearly a million refugees represents billions of dollars in lost wages, property, and economic stability.
Reported: The reports describe a massive volume of toxic and malicious content, including dehumanizing hate speech, calls for extermination, and pornographic images.
Directly caused: Facebook's algorithms hosted and amplified over 1,000 identified posts calling Muslims 'dogs', 'maggots', and calling for their extermination.
Indirectly caused: The spread of this content across millions of users in Myanmar, where Facebook is the primary source of information.
Inferred additional harm: Millions of users were exposed to state-sponsored military propaganda and hate speech, deeply poisoning the social fabric of the country.
Reported: The reports describe systematic discrimination and targeting of Muslims and Rohingya.
Directly caused: Facebook's failure to provide adequate Burmese-language moderation disproportionately exposed Burmese Muslims to unchecked hate speech and harassment compared to other language groups.
Indirectly caused: The military used the platform to systematically target the Rohingya minority, denying their identity and portraying them as illegal 'Bengali' invaders.
Inferred additional harm: Systemic exclusion, denial of citizenship, and segregation of the Rohingya population were reinforced and normalized by the platform's content.
Reported: The reports describe severe violations of human rights, including ethnic cleansing, forced migration, and denial of citizenship.
Indirectly caused: The platform facilitated a campaign of ethnic cleansing, displacing over 730,000 people and violating their fundamental rights to life, liberty, and security.
Inferred additional harm: Ongoing systemic human rights abuses, statelessness, and lack of legal recourse for over a million Rohingya refugees in Bangladesh camps.
Reported: The reports describe the erosion of democratic transition and the use of state surveillance and censorship.
Indirectly caused: The military manipulated public opinion to undermine civilian leaders like Aung San Suu Kyi, and the government drafted restrictive hate speech laws that critics fear will be used for censorship.
Inferred additional harm: The complete derailment of Myanmar's democratic transition and the consolidation of military influence through digital psychological warfare.
Reported: The reports describe doxxing, hacking, and state surveillance.
Indirectly caused: Prominent activists and individuals had their personal photos, relationships, and locations exposed (doxxed) on Facebook, leading to targeted threats.
Inferred additional harm: Widespread monitoring of civilian communications and profiles by domestic security services using information gathered from the platform.
Reported: The reports describe severe harassment, doxxing, death threats, and terror experienced by individuals.
Directly caused: Users directly received death threats, pornographic images, and targeted harassment via Facebook Messenger and comments.
Indirectly caused: The widespread dissemination of dehumanizing hate speech comparing Rohingya to dogs and maggots caused immense trauma, fear, and anxiety among the Muslim minority.
Inferred additional harm: Widespread psychological trauma, PTSD, and severe distress affected hundreds of thousands of refugees and displaced persons who fled the violence.
Reported: The reports describe widespread fabrication of news, doctored photos, and false rumors.
Directly caused: Facebook's platform allowed the viral spread of fabricated stories, such as fake rapes and false warnings of imminent 'jihad attacks' or Buddhist mobilizations.
Indirectly caused: The population's inability to distinguish real news from fake, leading to a complete breakdown of shared consensus reality.
Inferred additional harm: Long-term erosion of trust in media, institutions, and digital information across Myanmar and other developing markets.
The systematic exploitation of Meta's recommendation algorithms by the Myanmar military to incite ethnic cleansing represents a critical national security and human rights disaster. It highlights how algorithmic amplification of toxic content can be weaponized by state actors to destabilize society and commit mass atrocities.