[0:00]The release of the Epstein files has been presented as a major transparency effort, opening millions of pages of court records, investigative documents, flight logs, and witness statements
[0:11]linked to convicted pedophile Jeffrey Epstein. And the yearslong investigations into his crimes. Millions of documents were released to the public in several batches, but questions remain
[0:22]over how much material has yet to be disclosed. What has been released offers [music] a deeper look into Epstein's world and the network of powerful figures, associates, and institutions
[0:33]connected [music] to him. Some of the victims, I'm going to call them survivors, of his crimes have long called for greater transparency. But alongside those demands for disclosure came calls for caution. Survivors and
[0:44]their advocates wanted to ensure that anyone who wished to remain anonymous would be protected. [music] While authorities said survivors names and other identifying information had
[0:55]been redacted, concerns emerged that some released documents may still have contained information capable of identifying survivors or other private individuals. [music]
[1:07]The Department of Justice warned that some sensitive personal information may have been released unintentionally. At the same time, public attention became fixated on the names appearing
[1:17]throughout the files. And it's important to stress that being mentioned in the Epstein files is not in itself evidence of wrongdoing. So just how much material has been released? What information may
[1:29]have remained or remains exposed? And when concerns were raised, were documents corrected, updated, or withdrawn? My colleagues Julia Byer and Lewis Sanders from DW's investigative
[1:40]unit have spent months analyzing these releases and trying to answer those questions, and they're here to discuss their findings with me. Welcome to you both. It's really great to have you
[1:50]here. Um so I'll start with you Julia. As I've just mentioned um you spent months analyzing these releases. So what was the single most um surprising,
[2:01]shocking or you know significant finding um from your investigation? >> Mhm. Um so actually there were two
[2:10]things but um one that made us shaking our heads all the time was that there's up till today uh still uh information to identify survivors unredacted and
[2:22]accessible for the public uh in in the DOJ's uh library accessible. So I'm talking about not only names, full names like first name and last name, email
[2:34]addresses uh and even some uh images. So faces are still um unredacted and that was so surprising because throughout the
[2:44]last months not also that the DOJ um faced criticism about their failure of not following the law and uh rejecting properly um uh the survivors of sexual
[2:55]abuse but also survivors uh um themselves they came uh out public uh also explaining how much impact it had
[3:05]uh being exposed And then still we found from some of them the names still unredacted in the files. So I think that was like one special thing we were like
[3:16]that's incredible. Uh such a big failure. um some of the files uh because we had both data sets uh we scraped our own a couple of days later but also we had the original data set and we co
[3:28]compared both of them [snorts] um and we could see files for example from unredacted um or when fights where victim survivors were uh not redacted
[3:39]these files were completely removed some were re-uploaded so that was also an interesting aspect like us saying uh yeah we DOJ said we transparent and we
[3:50]also uh promise to protect uh survivors and they call them victims >> and they didn't. >> Wow. Yeah. I mean that's that's pretty shocking stuff what you're saying there.
[3:59]Was it What about for you Lewis? What was what was your kind of take? What was the most shocking surprising significant um thing from your from your analysis?
[4:10]I think uh I mean I agree with Julia on many of the things especially seeing you know when we first started we didn't know what to expect going into this uh these files we were still kind of
[4:22]figuring out what would be our angle what would be our story and as we got deeper into the files we started noticing that the victim's uh personally identified information uh was effectively unredacted in places
[4:34]>> despite the fact uh that this was supposed to be protected by law through the transparency act, but also the DOJ made promises too. So, they were supposed to uphold that law.
[4:45]>> Um, I think one of the other surprising aspects was the fact that, you know, I think there was a lot of newsrooms and investigative units that were using AI
[4:55]LLMs to do the searches uh of content. And um while we were almost tempted to do that, uh we decided that if we did
[5:06]that, we would end up possibly losing out on on information along the process. And so we decided to go manually through the documents. We kind of distilled it into subsets of data and started going
[5:17]through them manually. >> And I mean, it's it's surprising. Of course, there's a warning on the DOJ website saying, "Hey, you know, there's graphic content here." But when you're looking
[5:27]through different uh documents, whether it be uh flight logs, um bank statements, etc., then all of a sudden you come across uh graphic uh survivor
[5:38]testimony. I mean, it, you know, it jars you. >> Yeah. I mean, it's it sounds like a a massive task that you both um undertook. Um and I mean, from what you're saying, I think it's a pretty good moment
[5:49]actually to bring in um actually one of the survivors testimony in front of Congress. Uh that was Danny Benki. Um and we're going to just take a listen to it and then we're going to pick uh back
[6:01]up afterwards. All right. So let's uh let's play that now. >> In December before the document release, our attorneys submitted 350 victim's names to the Department of Justice as
[6:13]victim's names to be redacted prior to the release. In that first release, I found my name in two places. In that next release in January, my name appeared again, but this time the
[6:23]redactions were so far worse. Um, the files displayed not only my name, but my phone number, my former addresses, where I worked, and other identifying
[6:33]information. When my name appeared in the third file release, it became difficult to believe that this was not intentional. Despite my lawyer repeatedly contacting
[6:43]the DOJ seeking protection, my information continued to be exposed. And it's not just my name. It's the identifying information. These documents
[6:54]contain disturbing yet incomplete accounts of my abuse. They were available not only for the entire world to see, but my child, my students, my friends, my employers, my colleagues,
[7:06]and my family. It was humiliating. My immediate reaction is how did this happen? I mean, how is this allowed to happen? And and and what's what are the
[7:16]consequences then? And and how can those names be released or sorry, retracted? >> I mean, there's a lot of question marks there still. The US Department of
[7:26]Justice hasn't really been transparent about how all of these names ended up being released publicly. Um, they haven't been transparent about how the
[7:36]redaction process took place. And so at the moment it's it's unclear. They've only said, "Look, we admit that there was an error." They said this three days after the the latest uh release on
[7:47]January 30th. They said there a government lawyer basically said there was an error uh likely caused technical or human uh various factors involved. Um
[7:58]but it's one of those things you know once these names are out there even if they go through they start pulling down files which is what they did start further redacting them um those names
[8:08]are out there and the issue is though what we have seen also is that even despite the efforts by the DOJ to go through revise its redaction process
[8:18]flag cases uh and deal with them victims names are still out there even today. >> Yeah. And so this is this is a major issue. Now what's what's the outlook
[8:30]forward? So there are lawmakers that have recently introduced a new bill called the Epstein Transparency Act 2, which would effectively allow basically
[8:39]US state attorneys and victims to take the DOJ to court for the failures in the redaction process. I mean this, like she just said, I mean this is irrevocable
[8:50]damage that will carry on for the rest of her life. friends, colleagues, people she doesn't know, random people on the internet, all now know the things that she went to and what she testified.
[9:00]>> And Danny Benky's gone public, right? >> Essentially, so there are other people who have decided to just sort of stay quiet and just sort of even though their names information is out there, right?
[9:10]Uh you also went through audio files as well, right? Tell us about about what you found there. >> We collaborated with the Fhua Institute.
[9:18]Um it's called IDMT um um digital media technology and they they have a group that's uh specialized on uh detecting uh
[9:29]audio authenticity um manipulation and uh speech synthesis detection >> and they uh we shared with them uh over
[9:38]150 uh audio files and they were able with the uh state-of-the-art technology to um to analyze and to find out are
[9:48]they actually uh modified audios in there. Um, were there original audios the same like with documents like original audios and then later altered
[9:59]audios and were the audios where the same uh microphone setup was being used. Right? that's part of their technology and they found out um that uh the
[10:10]redection was inconsistent um in a lot of files audio files um which confirmed our um analysis with the documents as
[10:19]well and and images um and meaning what like what means uh inconsistent um of redection was on one side the style of
[10:28]redection. So sometimes there were parts uh beat and sometimes used a different sound. And then we found which is incredible in one in the same data set
[10:39]like published at the same moment and also still today both audio files. Um it's the same audio file but in one you
[10:49]can clearly hear it's an alleged witness that's uh saying I have information regarding the Epstein investigation and the person is uh saying his full name
[11:00]and his phone number and it's unredacted still public and the same audio file exists but his full name is beeped and his phone number as well. It's pretty
[11:10]shocking that like alleged witnesses of information. I I'm really flabbergasted by a lot of what you're you're telling me um today. I mean, this has been um a
[11:19]monthsl long process for for both of you, right? Um sounds like a lot of work. Sounds like it's pretty grueling. Louis, can you tell us just a little bit about that process? How that has been
[11:30]for you? >> I mean, in many ways, when you're dealing with content of child sexual abuse, it it's difficult. I mean it impacts you. It as you kind of we went
[11:42]through a lot of files. We were going through a lot of different things and it it can start warping a bit your sense or your perception of reality. You know when you see this whole entire network
[11:51]that effectively enabled uh this person and their Epstein and their associates uh to do what they did um you realize how extensive it was, how many players
[12:02]were involved. Um, yeah, I mean that that takes its toll on you. And one of the things that as we were going through it, Julia and I, we always made sure to kind of check in with each other. How
[12:14]are you doing today? Try to take breaks. Also, kind of separate the space that you have or time that you have when you're going through the files because you never know when it when something like that was going to come up,
[12:25]>> right? And uh and that was really important for us to check in with each other, just see where we're at. >> How are you doing today? Did you end up reading something? you want to talk about it and and that was huge. I think that that really helped us get through a
[12:36]lot of the work. >> Yeah, totally. >> Yeah. And now just tell us um if we want to read more or learn more about your work, where can we find it?
[12:45]>> Our report uh you can find this on Instagram and Tik Tok. >> Yeah. And we also publish on dw.com an online article which is uh the online
[12:56]article is focusing and explaining more how we did it, how we obtained the data, how did we verify it. Um that's a little bit more behind the scenes. >> Okay.
[13:06]>> Um and then uh you can also follow us on our social media accounts and and >> right >> get information. Yeah. >> Great. Well, it'll be um I'm I'm I'm
[13:17]going to read it for sure. I'll make sure to keep a close eye on um on this story. I don't feel like this is going to be the last time um unfortunately that we're going to be speaking about this. Um but really great to have both
[13:29]of you here, Julia Byer and Lewis Sanders from DW's investigative team. Thank you so much for joining us here in Berlin. And don't forget that if you want to learn more about their work, you
[13:39]can of course find it on the DW website, dw.com, and of course on social media too. I'm Pablo Foilas here in Berlin for me and the team. Thanks for watching and
[13:51]I'll see you again soon.