DOJ acknowledged 'human or technical error' exposed survivors' identities in January 2026 release of 800,000+ Epstein-related documents. DW's forensic analysis found dozens of files with unredacted names, faces, emails, and audio recordings of victims and witnesses still online.
DOJ's Epstein Files Still Expose Abuse Survivors Despite Redaction Promises
Once sensitive information is disclosed, the damage cannot be undone.
How did DW even know what was in the original release if the DOJ kept changing the files?
They found a public archive on GitHub that had captured the documents in the early days. By comparing that snapshot to what was on the DOJ website weeks later, they could see what had been removed, added, or modified. They used cryptographic hashing—basically a digital fingerprint—to verify which files matched the originals.
So the DOJ was actively taking files down and re-uploading them?
Yes, and not always in a way that suggested a coordinated effort to fix the problem. Some files were re-released with additional redactions, but others seemed to have been altered in inconsistent ways. It looked less like a systematic review and more like ad hoc changes.
What made the audio files particularly problematic?
The redactions were wildly inconsistent. In one file, a tipoff caller's phone number was bleeped out but their name was still audible. In another version of the same file, both the name and contact information were present. It suggested that witnesses and informants weren't being treated with the same care as victims.
Did the DOJ acknowledge any of this when DW confronted them?
No. DW presented their findings directly to the department, and the DOJ didn't respond. It wasn't until April that the Inspector General announced an audit. By then, the damage was already done—survivors were experiencing harassment and threats.
What's the legal recourse now?
In July, lawmakers introduced a second version of the Transparency Act that would allow survivors and state governments to actually sue the DOJ for non-compliance. The original law was supposed to protect survivors' identities, but there was no enforcement mechanism when the DOJ failed.
Can the information ever really be removed from the internet?
Not really. Once it's public, it's captured in archives, shared, downloaded. One survivor told Congress that the damage can't be undone. The information is still on public platforms and, in some cases, still on the DOJ website itself.
Il Polso
- DOJ released 800,000+ Epstein documents in January 2026; acknowledged 'human or technical error' exposed survivors' identities within three days
- DW investigation found dozens of files with unredacted names, faces, emails, and audio recordings still publicly accessible six months later
- More than 500,000 files were removed from DOJ website between initial release and DW's analysis weeks later
- Lawmakers introduced Epstein Files Transparency Act II in July to allow survivors to sue DOJ for non-compliance
DOJ acknowledged 'human or technical error' exposed survivors' identities in January 2026 release of 800,000+ Epstein-related documents. DW's forensic analysis found dozens of files with unredacted names, faces, emails, and audio recordings of victims and witnesses still online.
A DW investigation found that the DOJ's January release of Epstein documents still contains identifying information of abuse survivors despite promises to redact sensitive data. Hundreds of files with names, faces, and contact details remain publicly accessible, causing ongoing harm to victims.
In January, the Department of Justice released hundreds of thousands of documents related to Jeffrey Epstein, a financier convicted of orchestrating a sprawling network of sexual abuse. Within three days, the department acknowledged that the release had gone catastrophically wrong. Names, faces, email addresses, and other identifying information of survivors had been published online. A lawyer for the DOJ told a federal court that human error and technical failures had caused the breach, and the department promised to pull the documents, review them properly, and republish them with correct redactions.
Six months later, that promise remained unfulfilled. Deutsche Welle's investigative unit, working with its data and innovation teams, spent months analyzing the Epstein Library—the DOJ's official archive of released documents. What they found was damning: dozens of files still contained information that could identify survivors, witnesses, and informants. The investigation revealed not just carelessness, but a pattern of inconsistent and incomplete redaction work that suggested no systematic effort to protect the people whose lives had already been shattered.
The scale of the problem became clear only through painstaking technical work. By mid-February, DW had downloaded more than 800,000 files from the DOJ's archive. But the archive was constantly shifting—files were added, removed, and modified. To understand what had actually been released in January, DW obtained an earlier snapshot of the documents preserved by a public data archive on GitHub. Comparing the two versions revealed that more than 500,000 files had been removed from the DOJ website between the initial release and DW's download weeks later. Using cryptographic hashing—a technique that creates a unique digital fingerprint for each document—DW could verify which files matched the originals and which had been altered.
What emerged was a portrait of haphazard work. Some files had been re-released with additional redactions. Others had been modified in ways that suggested redactions had been added or removed. In one case, a survivor's testimony was redacted throughout an entire transcript until the final page, where a lawyer thanked her by name. That name remained unredacted. It was not an isolated mistake. DW found dozens of similar instances where survivors' identities had been left exposed.
The audio files presented a different kind of problem. DW collaborated with the Fraunhofer Institute for Digital Media Technology in Germany, whose forensic audio researchers used sophisticated tools to analyze more than 150 audio recordings of victim statements and tipoff calls. The researchers discovered that redaction styles varied wildly across files, and that some redactions were inconsistent or incomplete. In one recording, a person who had called an official tipoff hotline to report information about Epstein's abuse had their phone number bleeped out—but their name remained audible. In another version of the same file, both the name and the direct contact information were present. Witnesses and informants, it appeared, had not received the same protections as victims. "If one observes the content of these audio files, one can see that a lot of personal information has not been redacted," said Milica Gerhardt, a forensic audio researcher at the institute.
When DW presented its findings to the DOJ, the department did not respond. By April, however, the DOJ's Office of the Inspector General announced that it was launching an audit of the department's compliance with the Epstein Files Transparency Act, the law that was supposed to guarantee that survivors' personally identifiable information would be protected. In July, lawmakers introduced a second version of the act, which would allow survivors and state governments to sue the DOJ for failing to comply with the original law's requirements.
But for the survivors whose names and faces were already public, the damage was irreversible. Those whose information had been exposed reported harassment, humiliation, and physical threats. One survivor testified before Congress in May about the psychological toll. "I cannot live without looking over my shoulder," she said, describing how her personal information had been exposed. "I can only imagine the long-term impact this mistake will have on my life." She added, with bitter clarity: "While the rich and powerful remain protected by redactions, my name was exposed to the world." The information that had identified her remained available on public platforms and, in some cases, still on the DOJ website itself. The promise to fix the problem had not been kept, and the harm continued to accumulate.
Citazioni salienti
I cannot live without looking over my shoulder. I can only imagine the long-term impact this mistake will have on my life.— Sexual abuse survivor testifying to Congress in May
While the rich and powerful remain protected by redactions, my name was exposed to the world.— Same survivor, in testimony to Congress