“PDF Forgeries Are Surprisingly Rare”, by Gwern · Gwern.net

Skip to main content

Warning: JavaScript Disabled!

For support of key website features (link annotation popups/popovers & transclusions, collapsible sections, backlinks, tablesorting, image zooming, sidenotes etc.), you must enable JavaScript.

Site
Me
New
Blog
Links
Patreon
Substack

PDF Forgeries Are Surprisingly Rare

infosec, scientific bias

N/A

by: Gwern
2022-01-23–2025-03-19
finished
certainty: log
similar

One kind of fraud is striking in its absence online: tampered or forged PDFs. People create malicious videos, photos, chat logs, and Microsoft Word documents all the time to scam and propagandize people, or publish entire PDFs full of garbage science, but they don’t edit existing PDFs.

Once in a while I see someone object to a paper I host on Gwern.net by saying “but that’s not on a real journal website! it’s not peer-reviewed! it’s just some random asshole’s personal website! You can’t believe that!” Aside from bemusement over people not believing PDFs can exist elsewhere (do they not understand the idea of “files”? Apparently many young people struggle with it. Or are they just extremely eager to obey copyright law & can’t even imagine just copying a PDF?), it’s interesting how well this actually works.
There is plenty of incompetence, fraud, and malice online, often in PDFs… but only new PDFs. I can’t think of a single fraud accomplished by editing a real PDF & just uploading it for Google Scholar etc. or where I’ve been burned by even mislabeling.
You can just search for a paper title, download it, and trust that ~100% of the time, you are getting what you thought you were getting, with the main caveat being that you may be downloading the author’s draft or a preprint and not the finalized version (particularly in economics, where papers might go through many preprints, sometimes changing the results substantially along the way, and take anywhere up to a decade to reach final publication). And when you do find a PDF claiming something malicious, like claiming to use statistics to show that ‘Trump won the 2020 US presidential election’, it’s always a ‘new’ PDF, which is forthright about it being a new unpublished ‘white paper’ or somesuch, and doesn’t purport to be a published paper. Or if it was a forged or edited document, it was usually clearly exported from Microsoft Word or another word processor (eg. all the forgeries exposed by anachronistic use of the Calibri font).
Whereas, if you were so epistemically careless with images on, say, Facebook, you would wind up with a folder stuffed full of lying images which have been Photoshopped, claimed to be things other than what they are, ‘deep faked’, etc.

PDF forgery is striking because it’d be so easy to do: find a useful research paper, edit it in any of the many PDF utilities, upload anywhere, wait for people to copy it (as they do), then take down yours; now you have an authoritative peer-reviewed research paper floating around the Internet with no links to you, in the perfect crime. (Or better yet: upload it directly to Libgen/Sci-Hub and let everyone else redistribute it.)
Given how rarely people check the original papers, and how retracted studies like the Wakefield autism/vaccine study or blatant propaganda like Operation Denver documents will circulate among the epistemically-lazy indefinitely, a not-too-blatant forgery can get into widespread circulation for a long time before anyone notices. (In the rare cases anyone tampers with PDFs, it quickly turns into a technical morass and he-said-she-said, and requires several orders of magnitude more effort to prove than to do; consider the hoops epxperts had to jump through in the Craig Wright cases, or even just a landlord editing a contract—where the contract was done through a digital document timestamping service!)
And it’s not as if there are no zealots or fanatics or malefactors willing to do so—historically, scribes tamper with documents all the time! (“Written by Confucius” or “apropos of nothing, now I, Josephus the Jew, will tell you how wonderful Jesus Christ was”…)
Why can you just download PDFs off any random asshole’s website (like mine)?

Because there’s no Photoshop for PDFs, maybe? Places like arXiv provide TeX sources, but that’s still a dark art for most would-be forgers and fanatics and con artists. PDFs are not necessarily hard to edit, but editing PDFs is not part of a normal workflow for most people: even little kids will edit photos for social media, but outside big corporations & designers, PDFs are strictly “write-only” formats—your document system compiles source documents to PDFs, and you never edit the PDF, only the source documents. (It is somewhat analogous to the compiled binary of a program: a PDF is focused on laying out, pixel by pixel, how a printed document should look; it may not even contain the original text, much less any of the structure. Just as there are hackers who specialize in understanding and changing raw binary computer code, there are people who specialize in editing PDFs… but not many.)
This is enough to push malefactors into other approaches. After all, if editing photos can work so well, why bother with the much harder editing of PDFs?
As the joke goes, PDFs don’t need to outrun the (Russian?) bear, they just need to outrun the other format.
[Return to blog index]

Error: JavaScript disabled.

Backlinks, similar links, and the bibliography require JS enabled to load.

Similar Links
[Similar links by topic]

[ Send Anonymous Feedback ]

[Quote Of The Day]
[Site Of The Day]
[Annotation Of The Day]

[adblock public service announcement]

​

One form of online fraud that is noticeable is the creation of tampered or forged PDFs, which contrasts with the proliferation of malicious videos, photos, and chat logs used for scams and propaganda. The author observes that while there is considerable incompetence and malice online, mostly found in other media, the forgery of existing PDFs is surprisingly rare. The author notes that people often struggle to believe that files like PDFs can exist outside of their original contexts, suggesting that adherence to copyright or a lack of understanding regarding the concept of digital files acts as a barrier to some forms of fraud.

The author suggests that fraudulent PDFs are typically "new" documents, rather than edits of existing ones. For instance, when false claims, such as fabricated statistics regarding election results, are presented in PDF format, they are usually labeled as unpublished white papers or similar documents rather than purporting to be established, published research. In cases where documents are forged or edited, the alterations are often easily detectable, frequently exposing artifacts from editing software like Microsoft Word, such as inconsistent font usage.

The potential for PDF forgery as a "perfect crime" is noted, suggesting that it would be relatively simple to take a legitimate, peer-reviewed research paper, edit it using PDF utilities, and circulate it widely online without leaving a traceable link to the perpetrator. This is exacerbated by the inertia of the public, who rarely scrutinize original sources, allowing not-too-blatant forgeries to circulate for extended periods before being noticed. In instances where tampering does occur, the process often becomes a complex dispute requiring significant effort to prove, similar to historical legal cases involving digital documentation.

The difficulty in editing PDFs is attributed to the format itself; PDFs are not generally designed for casual editing. They function more like compiled binary files focused on the precise spatial layout of printed material rather than containing easily editable textual content or structure. Unlike image manipulation, which is widely accessible, editing PDFs is not part of a common workflow for most users, making it a less attractive target for malefactors than other methods. The author ultimately posits that because editing PDFs is a much harder endeavor, malicious actors often resort to other, less difficult forms of manipulation, such as editing photographs.