The Platform That Became a Weapon

When Facebook expanded into Myanmar in 2012, the country was emerging from decades of military junta rule and had just opened to foreign investment and mobile networks. Smartphone penetration went from near-zero to tens of millions of users in a few years. For the vast majority of those users, Facebook was not an app — it was the internet itself. The platform came preloaded on the cheapest handsets. There was no competing news ecosystem, no independent media infrastructure, no digital literacy tradition to contextualize what appeared in the feed.

The Tatmadaw — Myanmar's military — grasped the opportunity immediately. Military-controlled accounts seeded Facebook with fabricated stories: invented rapes, invented attacks, invented threats. Buddhist nationalist groups amplified coordinated hate speech portraying the Rohingya Muslim minority as an existential threat to the country. The UN Independent Fact-Finding Mission on Myanmar, which concluded its investigation in 2018, called Facebook "a useful instrument for spreading anti-Rohingya content" and said the platform "played a determining role in the violence that followed." UN Fact-Finding Mission

This was not a fringe platform problem. It was a population-scale information environment shaped, in the most algorithmically consequential sense, by a company headquartered in Menlo Park, California.

The warnings came early. Between 2012 and 2017, civil society organizations, researchers, and journalists repeatedly told Facebook that its platform was being used to incite violence against the Rohingya. Local activists documented specific posts. International human rights groups flagged coordinated hate speech campaigns. Facebook's response was to continue optimizing for growth in the market.

Sarah Wynn-Williams, Facebook's Director of Global Public Policy, was sent to Myanmar in 2012 to identify growth opportunities. In her 2025 memoir and congressional testimony, she described raising concerns about a Burmese-language content moderator being in cahoots with the junta — concerns that were dismissed. When she pushed to have Facebook's Community Standards translated into Burmese so users could understand the rules and moderators could enforce them, the communications team told her: Myanmar isn't a priority country. Courthouse News

On July 4, 2014 — while Facebook was still refusing to translate its own community standards into Burmese — a fabricated story about a Muslim man raping a Buddhist woman went viral on the platform. It set off riots. Mosques were attacked. People died. This was three years before the 2017 genocide. Facebook had been warned. The company's internal response was to continue optimizing for growth.

Amnesty International's 2022 investigation found that Facebook's algorithms did not merely fail to remove anti-Rohingya content — they actively spread it. The report concluded that Meta's systems proactively amplified and promoted content that incited hatred and violence against the Rohingya, beginning as early as 2012. The engagement-maximization algorithm discovered what anyone who has studied the psychology of online radicalization already knows: fear and hatred generate more clicks, more shares, more time-on-platform than neutral content. So the algorithm fed it back. Amnesty International

An internal Facebook document leaked as part of the Facebook Papers — the cache of internal documents provided to Congress by whistleblower Frances Haugen in 2021 — stated that the company had evidence that hate speech, divisive political speech, and misinformation on Facebook were affecting societies around the world, and that "core product mechanics such as virality, recommendations, and optimizing for engagement were a significant part of why these types of speech flourish on the platform." CNN

The same internal documents revealed that Facebook did not have Burmese-language content moderation capacity remotely proportional to its user base in the country. Myanmar was both a major growth market and a language Facebook had decided was not a priority.

The coordinated Tatmadaw offensive against the Rohingya in Rakhine State began on August 25, 2017. What followed has been documented by Médecins Sans Frontières, the United Nations, and multiple independent human rights organizations as a military campaign meeting the definition of genocide: systematic killing, sexual violence as a weapon of war, and the mass burning of villages.

Médecins Sans Frontières estimated at least 6,700 Rohingya killed in the first month alone — the period between August 25 and September 24, 2017. Rohingya survivor organizations put the total death toll higher: estimates range from 10,000 to more than 25,000. The UN Human Rights Council's Fact-Finding Mission documented mass graves, gang rape, and the deliberate burning of entire villages. UN Fact-Finding Mission

By the end of 2017, more than 730,000 Rohingya had fled to Cox's Bazar, Bangladesh, creating the world's largest refugee camp. Today, more than 1 million Rohingya remain in exile. The International Court of Justice is hearing a genocide case brought by The Gambia against Myanmar — hearings on the merits concluded in January 2026. ICJ No Meta executive has been charged with anything.

After the 2017 genocide, Meta commissioned an external human rights impact assessment. The company published it in November 2018. The report acknowledged that Facebook had been used to spread hate speech in Myanmar and that the company had not done enough to prevent its platform from being used to "foment division and incite offline violence." Meta HRIA 2018

This is Meta's own language in a report Meta commissioned and published. The acknowledgment was followed by commitments to improve. Amnesty International's 2022 investigation found that the algorithmic amplification of anti-Rohingya hate speech had in fact continued after the genocide — the engagement-maximization architecture remained intact. Amnesty International

In 2022, a coalition of organizations filed a complaint with the U.S. Securities and Exchange Commission alleging that Meta had misled shareholders by publicly claiming to protect human rights while internally knowing its systems were producing the opposite outcome in Myanmar. The complaint cited the Facebook Papers as evidence that executives were aware of the harm and chose not to act. Open Society Foundations Zero criminal charges have been filed against Meta, Mark Zuckerberg, or any company executive in any jurisdiction.

In 2021, Rohingya survivors and advocacy organizations filed a $150 billion class action lawsuit against Meta in US federal court. The complaint alleged that Facebook's algorithm and the company's failure to act on known warnings constituted legal liability for the genocide. Gizmodo

In 2024, a federal judge dismissed the case, citing the statute of limitations. The plaintiffs appealed. In April 2026, a Ninth Circuit panel dismissed the appeal under Section 230 of the Communications Decency Act, which provides broad civil immunity to online platforms for content posted by users. The ruling held that even if Facebook's algorithms proactively amplified hate speech, the platform remained protected. The Diplomat

This is the legal architecture that makes corporate accountability for genocide legally impossible in the United States. Section 230 was passed in 1996 to protect early internet forums from liability for user comments. It has become the shield that protects a trillion-dollar company from liability for a role in genocide that its own internal documents acknowledge.

The International Court of Justice case — The Gambia v. Myanmar — is directed at Myanmar's military, not Meta. Meta is not a defendant in any active case anywhere in the world. Mark Zuckerberg has never been charged, deposed, or held legally accountable in any jurisdiction for what happened in Myanmar. His net worth, as of 2025, is approximately $200 billion.

The Rohingya genocide was not caused by Facebook. But the UN, Amnesty International, and Meta's own commissioned researchers concluded that Facebook's systems amplified and accelerated a campaign that was already underway — and that the company had been warned, repeatedly, and chose growth. More than 1 million people are in exile. The algorithm is still running.