From the Twitter Files: Twitter, The FBI Subsidiary
The latest documents show, in bulk, the grotesque master-canine relationship between the FBI and Twitter
The Twitter Files has to be the craziest story in the history of journalism. There’s new drama every three minutes, it seems. The latest development had my phone blowing up with queries from multiple outlets. These included the New York Times and the Washington Post, two papers which didn’t call after the original story (although the Post, amusingly, did take time to temporarily label me a “conservative journalist”) but are suddenly hot for comment now.
The Twitter Files has to be the craziest story in the history of journalism. There’s new drama every three minutes, it seems. The latest development had my phone blowing up with queries from multiple outlets. These included the New York Times and the Washington Post, two papers which didn’t call after the original story (although the Post, amusingly, did take time to temporarily label me a “conservative journalist”) but are suddenly hot for comment now.
The new controversy is over the apparent banning of a series of high-profile journalists, including Middlebrow Retweet Specialist Aaron Rupar and the post-ESPN, post-MSNBC, Fangoria version of Keith Olbermann. I’m against the banning of journalists and will be happy to say so in this case once I’ve had time to look at it — I haven’t — but I’m beyond puzzled that media writers seem to think this has anything to do with me. I don’t work for Elon Musk, and I’m not his keeper. I’m just a journalist working a story, and the piece published today — “Twitter, The FBI Subsidiary” — I think moves the needle forward significantly on an unrelated, more important topic.
A few housekeeping notes. One, the weekly America This Week podcast with the incomparable Walter Kirn is coming, later tonight (we discuss the FBI story). Also, I’ll be posting today’s Twitter thread tomorrow morning (please excuse me for not having it ready tonight, I feel like I haven’t slept for a year). Lastly, I’m hoping sometime this weekend to have another story out, explaining in greater detail what we’re finding and what we think it means.
I started working on the “Twitter Files” story just over three weeks ago, after receiving an incredible invitation to examine “censorship and manipulation of opinion at a global level” at everyone’s favorite bird site.
After countless fits and starts, and enough plot twists to fill a dozen pulp novels, I’m finally comfortable enough reading the documents to see the outlines of the animal I came to hunt. Every day, the Twitter Files reveal more about how the machinery of state censorship works. We learn more and more every day about how the government collects, analyzes, and flags social media content in a neverending, cyclical process. The state isn’t a bit actor in a mostly-private “content moderation” movement. It’s the central player, clearly the boss of the whole operation, and clearly also the driving force in its expansion, a truth we can show in pictures.
I hope to flesh this out more in a coming piece, but a rough picture might show information traveling in a giant circle. The journey begins when rivers of “firehose” bulk data pour out of the back end of companies like Twitter and — either traveling directly or via middleman private contractors like Dataminr — end up in the possession of a smorgasbord of enforcement and intelligence agencies in Washington. There, these colossal data piles are analyzed for all sorts of predictable purposes, from tracking terror suspects to making economic forecasts.
We’ve long known about those analytical processes, but the Twitter Files show something new. We now have clear evidence that agencies like the FBI and the DHS are in the business of mass-analyzing social media activity — your tweets and mine, down to the smallest users with the least engagement — and are, themselves, mass-marking posts to be labeled, “bounced,” deleted or “visibility filtered” by firms like Twitter. The technical and personnel infrastructure for this effort is growing. As noted in the thread, the FBI’s social media-focused task force now has at least 80 agents, and is in constant contact with Twitter for all sorts of reasons.
The FBI is not doing this as part of any effort to build criminal cases. They’ve taken on this new authority unilaterally, as part of an apparently massive new effort to control and influence public opinion.
These agencies claim variously to be concerned about election integrity, foreign interference, medical misinformation, and monitoring domestic extremism, among other things. As crises wax and wane, the building out of the censorship infrastructure to ever-bigger and broader dimensions has been constant, suggesting that creating and deploying the tool to manipulate opinion was always the real end.
The Twitter Files don’t allow us to see inside the agencies. What we can do however is map the different points of re-entry at the end of the circle, when information returns to a company like Twitter flagged for moderation.
Is the request coming directly from the FBI or DHS? Is it coming from the government of a state like Colorado or California? Is it coming from a quasi-private entity like Stanford’s Election Integrity Project, a partner of the Center for Internet Security, which in turn is a partner of the DHS? Does a request arrive as a personal email sent to someone like very-former Deputy General Counsel Jim Baker, whose previous job involved being the FBI’s top lawyer?
We’ve seen government-flagged requests entering Twitter from at least a half-dozen routes, and we just started looking. We’ve seen flags on factually incorrect material, flags on material more controversial than wrong (DHS flagged a Trump tweet saying “Big problems and discrepancies with mail-in ballots all over the USA), and flags where the tweet was right and the government was just wrong. We’ve seen chatter where the FBI or DHS is cited openly, and other discussions where Twitter personnel refer mysteriously to the flags coming from parties variously called “our stakeholders,” “the escalators,” “partners,” and other names.
So far, the chief characteristic of the modern digital censorship regime seems to be the way public agencies and private contractors collaborate so smoothly on the project. Follow the journey of any particular piece of content, and you’ll find it may have moved in and out of the private sector multiple times on the way to being zapped at the finish line.
A lot of this was known before, but we’re seeing how it works at most every link of the chain now. It’s exciting, and I have every hope we’ll know twice as much by next week. Until then, thanks to all for hanging in there. This has been a difficult time in some ways, but it’s always a good sign when you get to see things — lots of things — agencies like the FBI would rather you didn’t. More TK. Have a good weekend, everyone.