Americans were left in shock as the first small bit of the Epstein files came out. Jokes about the level of redaction in the paper went wild on social media, a sort of shock-laugh at the horror revealed in the text documents… until someone realized the redaction wasn’t actually a permanent feature of the text box, rather something that could – pretty easily – be removed. Vulnerable data was exposed as a result, including info about victims that could identify them to their abusers. Presumably, this had a hand in further delays to planned releases.
Worse, video files seem to be coming out unplanned with increasingly slipshod redaction on shots in motion, and while people scouring the videos haven’t spotted felonious material, people are definitely identifiable by the contents of these videos given someone who knows them finds it. Modern internet makes that a far more likely reality than anyone in those clips wants to consider!
Failed redaction is actually pretty common. Generally speaking, redacted documents are copies of full documents with sensitive data removed. We only need to know what we need to know: a business’s email exchange that contains evidence of corruption and sensitive employee data can be redacted so the sensitive data is hidden from us, the people who are watching from the sidelines, so the employees don’t get doxxed as a result of the documents being released to the public as required by things like the Freedom of Information Act, or discovery during a trial. National security can also be a concern – if someone in a text exchange accidentally reveals the location of a supply building in Carolina containing high explosives, well, the public probably shouldn’t know that! But simply removing the entire paragraph might alter the structure of the text in a way that makes it confusing, misleading, or ugly. Retyping has not historically been an option – one page would be easy, yes, but dozens, or hundreds? And doing so accurately, without any typos or slips that could possibly change the meaning of a given sentence that might be incriminating? The classic Sharpie-style black bars were the easiest way to remove data on physical paper, and only recently has the problem of fully digital – from generation to redaction – documents reared its head.
Making a permanent change to a document is tougher than you would expect, as metadata stores a heck of a lot of info about the document behind the scenes, and redaction itself doesn’t tend to delete information, it just covers it. Removing the data altogether is an option, but it’s not one that’s exercised often because it can affect the formatting of the document, which may be relevant to the rest of the information we are allowed to see in it. However, as a result, these redactions can fail. Black boxes in Microsoft are moveable. Cropped photos can be uncropped. Redacting something permanently would suck really bad in most consumer software, which is why most consumer software doesn’t let you do it, even on purpose! So, it’s just not suitable for the task of redacting because it was designed that way. How often have you cropped a picture in Word and then wished you hadn’t? Once?
How did this even happen? Rumors (which may be wrong) say DOGE cut the budget of the department doing the redaction, leaving them with a lower tier of software than they needed to complete the task as required. Maybe, because DOGE was too cheap to pay for an official service, the redaction failed. That’s not the only problem, however. As victims note, some pages have first names, and others have last names, and as a result, if you’re reading these together you can identify the person being named. It’s almost as if the pages were divided among people who were not instructed in exactly what needed to happen and who each had to try their best to comply. Other rumors suggest this terrible job is done on purpose, because the contents of these files are so horrific that the people tasked with covering up can’t bear to let the offenders hide. Whether it was staff, budget, time, or soft mutiny, some of these mistakes are so incompetent you’d think it was baby’s first redaction.
Context
Sometimes art programs store images in layers. Sometimes checking a PDF for redactions means making the redactions not permanent until publish. With these two problems in mind, mistakes like not merging layers, or using a program that doesn’t actually remove or block the text (as in, you can still copy it from behind the box) are somewhat understandable. That doesn’t mean it’s not a huge mistake. Redaction is there for a reason! There could be serious extra-legal consequences for the people who were brave enough to come forward and accuse these very wealthy and very powerful people of harming them.
Additionally, government offices trying to use Adobe to redact stuff keep running into the same problem where the text is still copy/pasteable behind whatever black or white bars they’ve tried to use to redact text, and something similar seems to have happened with the initial release a few weeks ago as of the writing of this article. Users could simply copy the text behind the box like the box wasn’t even there. The text has to stay intact, trying to actually destroy or scramble it is effectively impossible with the scale of the information being released here, but surely, bothering to train employees on using Adobe or whatever correctly would have spared some of the embarrassment of getting caught out like this more than once in the same scandal! It was like putting a vector box over the text in Word.
Video is much tougher: to redact video, one has to add a black box over the subject in every frame of the video they appear in, and at 30 frames a second (sometimes 60 on consumer cameras) this can take a looong time if the camera or the subject are moving a lot.
In many of these videos, people are dancing, or the person filming is panning the camera around a party; in many of these videos, the black boxes can’t keep up with the movement, and people who were behind black boxes become visible, to the point that the back boxes are acting like a big arrow saying “pay attention to this guy!” as stuff jerks around onscreen.
Some people who have legal experience argue that the files shouldn’t have been released (notably, the New York Times’ article on the matter has an absolutely terrible headline for what the article actually says!) because they did such a truly horrible job of protecting the people in the files with those terrible redactions. New people who were involved with Epstein have been revealed (including an exchange involving Elon Musk not getting invited to a party despite asking), but half the files, videos, and other assorted bits and bobs may never see the light of day because of how poorly this batch went (of course not including the videos and files of abusive material which can never be released). Freedom of Information worked out alright
