No, The FBI Is NOT ‘Paying Twitter To Censor’

Sigh.

Look. I want to stop writing about Twitter. I want to write about lots of other stuff. I have a huge list of other stories that I’m trying to get through, but then Elon Musk does something dumb again, and people run wild with it, and (for reasons that perplex me) much of the media either run with what Musk said, or just ignore it completely. But Musk is either deliberately lying about stuff or too ignorant to understand what he’s talking about, and I don’t know which is which, though neither is a good look.

Today, his argument is that “the FBI has been paying Twitter to censor,” and he suggests this is a big scandal.

This would be a big scandal if true. But, it’s not. It’s just flat out wrong.

As with pretty much every one of these misleading statements regarding the very Twitter that he runs, where people (I guess maybe just former people) could explain to him why he’s wrong, it takes way more time and details to explain why he’s wrong than for him to push out these misleading lines that will now be taken as fact.

But, since at least some of us still believe in facts and truth, let’s walk through this.

First up, we already did a huge, long debunker on the idea that the FBI (or any government entity) was in any way involved in the Twitter decision to block links to the Hunter Biden laptop story. Most of the people who believed that have either ignored that there was no evidence to support it, or have simply moved on to this new lie, suggesting that “the FBI” was “sending lists” to Twitter of people to censor.

The problem is that, once again, that’s not what “the Twitter Files” show, even as the reporters working on it — Matt Taibbi, Bari Weiss, and Michael Shellenberger — either don’t understand what they’re looking at or are deliberately misrepresenting it. I’m no fan of the FBI, and have spent much of the two and a half decades here at Techdirt criticizing it. But… there’s literally no scandal here (or if there is one, it’s something entirely different, which we’ll get to at the end of the article).

What the files show is that the FBI would occasionally (not very often, frankly) using reporting tools to alert Twitter to accounts that potentially violated Twitter’s rules. When the FBI did so, it was pretty clear that it was just flagging these accounts for Twitter to review, and had no expectation that the company would or would not do anything about it. In fact, they are explicit in their email that the accounts “may potentially constitute violations of Twitter’s Terms of Service” and that Twitter can take “any action or inaction deemed appropriate within Twitter policy.”

That is not a demand. There is no coercion associated with the email, and it certainly appears that Twitter frequently rejected these flags from the US government. Twitter’s most recent transparency report lists all of the “legal demands” the company received for content removals in the US, and its compliance rate is 40.6%. In other words, it complied with well under half of any demands for data removal from the government.

Indeed, even as presented (repeatedly) by Taibbi and Shellenberger as if it’s proof that Twitter closely cooperated with the FBI, over and over again if you read the actual screenshots, it shows Twitter (rightly!) pushing back on the FBI. Here, for example, Michael Shellenberger, shows Twitter’s Yoel Roth rejecting a request from the FBI to share information, saying they need to take the proper legal steps to request that info (depending on the situation, likely getting a judge to approve the request):

Now, we could have an interesting discussion (and I actually do think it’s an interesting discussion) about whether or not the government should be flagging accounts to review as terms of service violations. Right now, anyone can do this. You or I can go on Twitter and if we see something that we think violates a content policy, we can flag it for Twitter to review. Twitter than will review the content and determine whether or not it’s violative, and then decide what the remedy should be if it is.

That opens up an interesting question in general: should government officials and entities also be allowed to do the same type of flagging? Considering that anyone else can do it, and the company still reviews against its own terms of service and (importantly) feels free to reject those requests when they do not appear to violate the terms, I’m hard pressed to see the problem here on its own.

If there were evidence that there was some pressure, coercion, or compulsion for the company to comply with the government requests, that would be a different story. But, to date, there remains none (at least in the US).

As for the accounts that were flagged, from everything revealed to date in the Twitter Files, it mostly appears to be accounts that were telling a certain segment of the population (sometimes Republicans, sometimes Democrats) to vote on Wednesday, the day after Election Day, rather than Tuesday. Twitter had announced long before the election that any such tweets would violate policy. It does appear that a number of those tweets were meant as jokes, but as is the nature of content moderation, it’s difficult to tell what’s a joke from what’s not a joke, and quite frequently malicious actors will try to hide behind “but I was only joking…” when fighting back against an enforcement action. So, under that context, a flat “do not suggest people vote the day after Election Day” rule seems reasonable.

Given all that, to date, the only “evidence” that people can look at regarding “the FBI sent a list to censor” is that the FBI flagged (just as your or I could flag) accounts that were pretty clearly violating Twitter policies in a way that could undermine the US election, and left it entirely up to Twitter to decide what to do about it — and Twitter chose to listen to some requests and ignore others.

That doesn’t seem so bad in context, does it? It actually kinda seems like the sort of thing people would want the FBI to do to support election integrity.

But the payments!

So, there’s no evidence of censorship. But what about these payments? Well, that’s Musk’s hand-chosen reporters, Musk himself, and his fans totally misunderstanding some very basic stuff that any serious reporter with knowledge of the law would not mess up. Here’s Shellenberger’s tweet from yesterday that has spun up this new false argument:

That’s Shellenberger saying:

The FBI’s influence campaign may have been helped by the fact that it was paying Twitter millions of dollars for its staff time.

“I am happy to report we have collected $3,415,323 since October 2019!” reports an associate of Jim Baker in early 2021.

But this is a misreading/misunderstanding of how things work. This had nothing to do with any “influence campaign.” The law already says that if the FBI is legally requesting information for an investigation under a number of different legal authorities, the companies receiving those requests can be reimbursed for fulfilling them.

(a)Payment.—

Except as otherwise provided in subsection (c), a governmental entity obtaining the contents of communications, records, or other information under section 2702, 2703, or 2704 of this title shall pay to the person or entity assembling or providing such information a fee for reimbursement for such costs as are reasonably necessary and which have been directly incurred in searching for, assembling, reproducing, or otherwise providing such information. Such reimbursable costs shall include any costs due to necessary disruption of normal operations of any electronic communication service or remote computing service in which such information may be stored.

But note what this is limited to. These are investigatory requests for information, or so called 2703(d) requests, which require a court order.

Now, there are reasons to be concerned about the 2703(d) program. I mean, going back to 2013, when it was revealed that the 2703(d) program was abused as part of an interpretation of the Patriot Act to allow the DOJ/NSA to collect data secretly from companies, we’ve highlighted the many problems with the program.

So, by the way, did old Twitter. More than a decade ago, Twitter went to court to challenge the claim that a Twitter user had no standing to challenge a 2703(d) order. Unfortunately, Twitter lost and the feds are still allowed to use these orders (which, again, require a judge to sign off on them).

I do think it remains a scandal the way that 2703(d) orders work, and the inability of users to push back on them. But that is the law. And it has literally nothing whatsoever to do with “censorship” requests. It is entirely about investigations by the FBI into Twitter users based on evidence of a crime. If you want, you can read the DOJ’s own guidelines regarding what they can request under 2703(d).

Looking at that, you can see that if they can get a 2703(d) order (again, signed by a judge) they can seek to obtain subscriber info, transaction records, retrieved communications, and unretrieved communications stored for more than 180 days (in the past, we’ve long complained about the whole 180 days thing, but that’s another issue).

You know what’s not on that list? “Censoring people.” It’s just not a thing. The reimbursement that is talked about in that email is about complying with these information production orders that have been reviewed and signed by a judge.

It’s got nothing at all to do with “censorship demands.” And yet Musk and friends are going hog wild pushing this utter nonsense.

Meanwhile, Twitter’s own transparency report again already reveals data on these orders as part of its “data information requests” list, where it shows that in the latest period reported (second half of 2021) it received 2.3k requests specifying 11.3k accounts, and complied with 69% of the requests.

This was actually down a bit from 2020. But since the period the email covers is from 2019 through 2020, you can see that there were a fair number of information requests from the FBI:

Given all that, it looks like there were probably in the range of 8,000 requests for information, covering who knows how many accounts, that Twitter had to comply with. And so the $3 million reimbursement seems pretty reasonable, assuming you would need a decent sized skilled team to review the orders, collect the information, and respond appropriately.

If there’s any scandal at all, it remains the lack of more detailed transparency about the (d) orders, or the ability of companies like Twitter to have standing to challenge them on behalf of users. Also, there are reasonable arguments for why judges are too quick to approve (d) orders as valid under the 4th Amendment.

But literally none of that is “the FBI paid Twitter to censor people.”

And yet, here’s Elon.