---
title: "Xenophobia & Violence"
description: "From Stormfront to Myanmar. The infrastructure for online hate predates social media by a decade."
url: https://attentionmachine.org/en/symptoms/xenophobia/
lang: en
updated: 2026-10-25
license: "CC BY-NC-SA 4.0 (https://creativecommons.org/licenses/by-nc-sa/4.0/)"
---

# Xenophobia & Violence

The connection between online hate and real-world violence was documented before many people had heard of Facebook.

## The timeline

Stormfront — the first major hate site on the internet —[^src-southern-poverty-1995] was set up in 1995[^src-southern-poverty-2015]. By the early 2000s, researchers were studying[^src-glaser-2002] how far people in white racist chat rooms would go in advocating violence. In June 2004, the OSCE held a formal intergovernmental meeting in Paris[^src-osce-2004] specifically on racist and xenophobic online propaganda and its connection to hate crimes — nearly twenty years before the EU Digital Services Act[^src-regulation-2022] required platforms to assess this risk. By early 2007, the Southern Poverty Law Center[^src-southern-poverty-2007] counted roughly 12,000 white-supremacist propaganda videos, hate-rock concert reels and Holocaust-denial pseudo-documentaries openly available on YouTube and other video-sharing sites.

This is not hindsight. These warnings were contemporaneous, specific, and well-evidenced — documented before the platforms that amplified the problem at scale had become dominant.

## Myanmar: the documented case

Facebook's dominance in Myanmar began taking root around 2012[^src-stecklow-2018], as the country opened up after decades of military rule; by 2018 it had 18 million users, about a third of the population, and for most of the people online[^src-bsr-2018], Facebook was the internet. UN investigators[^src-independent-fact-finding-2018] and journalists documented Facebook's role in the amplification of anti-Rohingya hate speech from 2016 onward. Facebook itself acknowledged in 2018[^src-facebook-2018] that it had not done enough to prevent the platform being used for incitement to violence. In 2017, the Myanmar military conducted a campaign against the Rohingya people that the UN's Special Rapporteur on Myanmar said in March 2018[^src-ohchr-2018] bore "the hallmarks of genocide"; the UN Fact-Finding Mission later called for senior military officials to be investigated and prosecuted for genocide.

The causal pathway from Facebook amplification to atrocity is complex — existing ethnic tensions, military action, historical grievances all played roles. What is documented: Facebook was the primary information environment, it amplified content that dehumanised the Rohingya population, the UN Independent Fact-Finding Mission[^src-independent-fact-finding-2018] found the company's response to the hate speech on its platform "slow and ineffective", and the human-rights impact assessment Facebook itself commissioned from BSR[^src-bsr-2018] found that the platform was "being used by bad actors to spread hate speech, incite violence, and coordinate harm".

> **Dehumanisation**
>
> Representing a group as less than fully human — as vermin, disease, or invaders. It is not just insult: psychologically, it loosens the moral restraints that normally stop people from accepting or committing violence against others. Genocide scholars treat escalating dehumanising language as a recognised warning stage, which is why its mass amplification in Myanmar is documented so carefully rather than dismissed as mere offensive speech.
>
> **Sources**
>
> - Stanton, G.H. (1996, revised), [_The Ten Stages of Genocide_](https://www.genocidewatch.com/tenstages), Genocide Watch — stage four, "Dehumanization, when perpetrators call their victims rats, or cockroaches, cancer, or disease"; the stages are "predictable but not inexorable".
> - [UN Independent Fact-Finding Mission on Myanmar](https://documents.un.org/doc/undoc/gen/g18/274/54/pdf/g1827454.pdf) (12 September 2018), A/HRC/39/64 — "Dehumanizing and stigmatizing language against the Rohingya" (¶73); Facebook "a useful instrument for those seeking to spread hate" (¶74).

## The algorithm does not choose hate. It chooses engagement.

The mechanism is the same throughout this site. The algorithm does not have views on ethnicity or politics. It surfaces content that generates [strong emotional responses](https://attentionmachine.org/en/attention-machine/how-it-works/) — and hatred, fear, and tribal identity are [among the most engaging emotions](https://attentionmachine.org/en/attention-machine/your-brain/). What follows from that is narrower than the popular story: the platform lowers the cost of finding extreme content, and of finding each other, and it serves extreme material preferentially to people already leaning that way. Whether the recommendation engine drives the escalation on its own is [contested](https://attentionmachine.org/en/symptoms/radicalisation/).

An internal Facebook presentation from 2016[^src-horwitz-2020], studying political groups in Germany, found that 64% of all joins to extremist groups came through the platform's own recommendation tools — "our recommendation systems grow the problem," as the slide put it. The mechanism is documented on [How It Works](https://attentionmachine.org/en/attention-machine/how-it-works/).

## What the research shows on online-offline violence connections

Research on the relationship between online hate speech and offline violence finds correlations. A study by Müller and Schwarz[^src-muller-2021] found a significant positive relationship between Facebook posts with anti-refugee sentiment and attacks on refugees in Germany, at the local level — and, using internet outages as a natural experiment, showed the relationship weakened when access was cut.

The honest framing: a causal pathway has been documented, with rigour, in one country and period. It is not deterministic — most people who consume hateful content do not commit violence. The direction is consistent. The scale of the amplification is what changes the base rates.

**How we know — the natural experiment behind the anti-refugee finding**

The standard objection to "online hate correlates with offline attacks" is that towns with more hostility might simply have both more hateful posts and more attacks, with the internet just along for the ride. Müller and Schwarz (2021)[^src-muller-2021] addressed this with a _natural experiment_: they used local internet and Facebook _outages_ — disruptions unrelated to local attitudes — as a lever. Where outages cut access, the relationship between anti-refugee sentiment and attacks on refugees weakened. Because the outages were essentially random with respect to local politics, this gets much closer to a causal claim than a simple correlation.

**What it shows and its limits.** It is strong evidence that the platform _amplified_ existing sentiment into action in a specific country and period, not a universal constant. It does not say everyone exposed becomes violent — the claim is about shifted base rates across many communities, which is exactly how amplification at scale would look.

## Sources

[^src-osce-2004]: OSCE (16–17 June 2004), [_Meeting on the relationship between racist, xenophobic and anti-Semitic propaganda on the Internet and hate crimes_](https://www.osce.org/node/66048), Paris.

[^src-southern-poverty-2015]: Southern Poverty Law Center (10 June 2015), [_20 Years of Hate_](https://www.splcenter.org/resources/reports/20-years-hate/), Intelligence Report — Stormfront opened in March 1995, “the first major hate site on the Internet”.

[^src-southern-poverty-2007]: Southern Poverty Law Center (Spring 2007), [_Neo-Nazi Groups Share Hate Via YouTube_](https://splcenter.org/fighting-hate/intelligence-report/2007/neo-nazi-groups-share-hate-youtube), Intelligence Report 125 — the \~12,000-video count.

[^src-independent-fact-finding-2018]: [UN Independent Fact-Finding Mission on Myanmar](https://www.ohchr.org/en/hr-bodies/hrc/myanmar-ffm/index) (2018) — [A/HRC/39/64](https://documents.un.org/doc/undoc/gen/g18/274/54/pdf/g1827454.pdf).

[^src-facebook-2018]: Facebook (5 November 2018), [_An Independent Assessment of the Human Rights Impact of Facebook in Myanmar_](https://about.fb.com/news/2018/11/myanmar-hria/) — "we weren't doing enough to help prevent our platform from being used to foment division and incite offline violence."

[^src-bsr-2018]: BSR (October 2018), [_Human Rights Impact Assessment: Facebook in Myanmar_](https://about.fb.com/wp-content/uploads/2018/11/bsr-facebook-myanmar-hria_final.pdf) — the assessment Facebook commissioned.

[^src-stecklow-2018]: Stecklow, S. (15 August 2018), [_Why Facebook is losing the war on hate speech in Myanmar_](https://www.reuters.com/investigates/special-report/myanmar-facebook-hate/), Reuters special report — more than 1,000 posts attacking the Rohingya; 18 million Facebook users, about a third of the population.

[^src-ohchr-2018]: OHCHR (12 March 2018), [_Myanmar: UN expert calls for accountability over violence in Rakhine State_](https://www.ohchr.org/en/press-releases/2018/03/myanmar-un-expert-calls-accountability-over-violence-rakhine-state) — Special Rapporteur Yanghee Lee to the Human Rights Council.

[^src-southern-poverty-1995]: Southern Poverty Law Center, [_Stormfront_](https://www.splcenter.org/resources/extremist-files/stormfront/), Extremist Files — created by Don Black in March 1995, "the first major hate site on the Internet".

[^src-glaser-2002]: Glaser, J., Dixit, J. & Green, D.P. (2002), [_Studying Hate Crime with the Internet: What Makes Racists Advocate Racial Violence?_](https://doi.org/10.1111/1540-4560.00255), Journal of Social Issues 58(1).

[^src-horwitz-2020]: Horwitz, J. & Seetharaman, D. (26 May 2020), [_Facebook Executives Shut Down Efforts to Make the Site Less Divisive_](https://www.wsj.com/articles/facebook-knows-it-encourages-division-top-executives-nixed-solutions-11590507499), Wall Street Journal — the 2016 "64%" presentation.

[^src-muller-2021]: Müller, K. & Schwarz, C. (2021), [_Fanning the Flames of Hate: Social Media and Hate Crime_](https://doi.org/10.1093/jeea/jvaa045), Journal of the European Economic Association 19(4) ([working paper, 2019](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3082972)) — Facebook use and anti-refugee attacks in Germany.

[^src-regulation-2022]: [Regulation (EU) 2022/2065 (Digital Services Act)](https://eur-lex.europa.eu/eli/reg/2022/2065/oj), Articles 34 and 38 — systemic-risk assessment of algorithmic systems, and the non-profiling recommender option.

[What you can do →](https://attentionmachine.org/en/educate/what-to-do/)\
[Elections & Democratic Harm →](https://attentionmachine.org/en/symptoms/elections/)\
[Yes, but… →](https://attentionmachine.org/en/objections/)

---

Some content on this site was created with the assistance of AI tools and may contain errors. [How we use AI →](https://attentionmachine.org/en/about/how-we-use-ai/)