I have subscriptions to several writers on Substacke, Matt Taibbi among them. This came out yesterday.
This one isn't (shouldn't be) paywalled:
https://www.racket.news/p/financial...tle&isFreemail=true&r=1b256w&utm_medium=email
A few weeks ago, Ohio congressman and Judiciary Committee chairman Jim Jordan’s office
released a
letter to Noah Bishoff, the former director of the Financial Crimes Enforcement Network, or FinCEN, an arm of the Treasury Department. Jordan’s team was asking Bishoff for answers about why FinCEN had “distributed slides, prepared by a financial institution,” detailing how other private companies might use MCC transaction codes to “detect customers whose transactions may reflect ‘potential active shooters.’” The slide suggested the “financial company” was sorting for terms like “Trump” and “MAGA,” and watching for purchases of small arms and sporting goods, or purchases in places like pawn shops or Cabela’s, to identify financial threats.
Jordan’s letter to Bishoff went on:
According to this analysis, FinCEN warned financial institutions of “extremism” indicators that include “transportation charges, such as bus tickets, rental cars, or plane tickets, for travel to areas with no apparent purpose,” or “the purchase of books (including religious texts) and subscriptions to other media containing extremist views.”
During the Twitter Files, we searched for snapshots of the company’s denylist algorithms, i.e. whatever rules the platform was using to deamplify or remove users. We knew they had them, because they were alluded to often in documents (a
report on the denylist is_Russian, which included Jill Stein and Julian Assange, was one example). However, we never found anything like the snapshot Jordan’s team just published:
The highlighted portion shows how algorithmic analysis works in financial surveillance. First compile a list of naughty behaviors, in the form of MCC codes for guns, sporting goods, and pawn shops. Then, create rules: $2,500 worth of transactions in the forbidden codes, or a number showing that more than 50% of the customer’s transactions are the wrong kind, might trigger a response. The Committee wasn’t able to specify what the responses were in this instance, but from previous experience covering anti-money-laundering (AML) techniques
at banks like HSBC, a good guess would be generation of something like Suspcious Activity Reports, which can lead to a customer being debanked.
If Facebook, Twitter, and Google have already shown a tendency toward wide-scale monitoring of speech and the use of subtle levers to apply pressure on attitudes, financial companies can use records of transactions to penetrate individual behaviors far more deeply. Especially if enhanced by AI, a financial history can give almost any institution an immediate, unpleasantly accurate outline of anyone’s life, habits, and secrets. Worse, they can couple that picture with a powerful disciplinary lever, in the form of the threat of closed accounts or reduced access to payment services or credit. Jordan’s slide is a picture of the birth of the political credit score.
There’s more coming on this, and other articles forthcoming (readers who’ve noticed it’s been quiet around here will soon find out why). While the world falls to pieces over Tucker, Putin, and Ukraine, don’t overlook this horror movie. If banks and the Treasury are playing the same domestic spy game that Twitter and Facebook have been playing with the FBI, tales like the
frozen finances of protesting Canadian truckers won’t be novelties for long. As is the case with speech, where huge populations have learned to internalize censorship rules almost overnight, we may soon have to learn the hard way that even though some behaviors aren’t illegal, they can still be punished with great effectiveness, in a
Terminator-like world where computers won’t miss anything that moves.
What a crazy time we live in! See you from the Nevada caucus, and watch this space for other news soon.
