Meta in Myanmar (full series)

Between July and October of this year, I did a lot of reading and writing about the role of Meta and Facebook—and the internet more broadly—in the genocide of the Rohingya people in Myanmar. The posts below are what emerged from that work.

The format is a bit idiosyncratic, but what I’ve tried to produce here is ultimately a longform cultural-technical incident report. It’s written for people working on and thinking about (and using and wrestling with) new social networks and systems. I’m a big believer in each person contributing in ways that accord with their own skills. I’m a writer and researcher and community nerd, rather than a developer, so this is my contribution.

More than anything, I hope it helps.

Meta in Myanmar, Part I: The Setup (September 28, 2023, 10,900 words)

Myanmar got the internet late and all at once, and mostly via Meta. A brisk pass through Myanmar’s early experience coming online and all the benefits—and, increasingly, troubles—connectivity brought, especially to the Rohingya ethnic minority, which was targeted by massive, highly organized hate campaigns.

Something I didn’t know going in is how many people warned Meta‚ and in how much detail, and for how many years. This post captures as many of those warnings as I could fit in.

Meta in Myanmar, Part II: The Crisis (September 30, 2023, 10,200 words)

Instead of heeding the warnings that continued to pour in from Myanmar, Meta doubled down on connectivity—and rolled out a program that razed Myanmar’s online news ecosystem and replaced it with inflammatory clickbait. What happened after that was the worst thing that people can do to one another.

Also: more of the details of the total collapse of content moderation and the systematic gaming of algorithmic acceleration to boost violence-inciting and genocidal messages.

Meta in Myanmar, Part III: The Inside View (October 6, 2023, 12,500 words)

Using whistleblower disclosures and interviews, this post looks at what Meta knew (so much) and when (for a long time) and how they handled inbound information that suggests that Facebook was being used to do harm (they shoved it to the margins).

This post introduces an element of the Myanmar tragedy that turns out to have echoes all over the planet, which is the coordinated covert influence campaigns that have both secretly and openly parasitized Facebook to wreak havoc.

I also get into a specific and I think illustrative way that Meta continues to deceive politicians and media organizations about their terrible content moderation performance, and look at their record in Myanmar in the years after the Rohingya genocide.

Meta in Myanmar, Part IV: Only Connect (October 13, 2023, 8,600 words)

Starting with the recommendations of Burmese civil-society organizations and individuals plus the concerns of trust and safety practitioners who’ve studied large-scale hate campaigns and influence operations, I look at a handful of the threats that I think cross over from centralized platforms to rapidly growing new-school decentralized and federated networks like Mastodon/the fediverse and Bluesky—in potentially very dangerous ways.

It may be tempting to take this last substantial piece as the one to read if you don’t have much time, but I would recommend picking literally any of the others instead—my concluding remarks here are not intended to stand alone.

Meta Meta (September 28, 2023, 2,000 words)

I also wrote a short post about my approach, language, citations, and corrections. That brings the total word to about 44,000.


Above all, all my thanks go to the people of the Myanmar Internet Project and its constituent organizations.

Thanks additionally to the various individuals on the backchannel whom I won’t name but hugely appreciate, to Adrianna Tan and Dr. Fancypants, Esq., to all the folks on Mastodon who helped me find answers to questions, and to the many people who wrote in with thoughts, corrections, and dozens of typos. All mistakes are extremely mine.

Many thanks also to the friends and strangers who helped me find information, asked about the work, read it, and helped it find readers in the world. Writing and publishing something like this as an independent writer and researcher is weird and challenging, especially in a moment when our networks are in disarray and lots of us are just trying to figure out where our next job will come from.

Without your help, this would have just disappeared, and I’m grateful to every person who reads it and/or passes it along.

Thanks” is a deeply inadequate thing to say to my partner, Peter Richardson, who read multiple drafts of everything and supported me through some challenging days in my 40,000-words-in-two-weeks publishing schedule, and especially the months of fairly ghastly work that preceded it. But as ever, thank you, Peter.

16 October 2023