How to Address Unprocessable PDF Submissions in SafeAssign for Plagiarism Checks?

Question

I am an astronomy professor who recently started using Blackboard's SafeAssign tool to check for plagiarism in student submissions. I've encountered a persistent issue with one of my students, whose submitted PDF files consistently fail to be processed by SafeAssign. The error message displayed is:

SafeAssign Submission

Unprocessed Attachments

Essay.pdf

We are unable to process this file. If this is a supported file type, please try again later.

Despite these files appearing normal (i.e., readable in Blackboard and when downloaded), they cannot be processed by plagiarism checking sites without manual OCR, and even then, the hyperlinks are missing. This discrepancy is notable, especially considering the student’s poor performance on exams relative to their online submissions. I suspect this might be a deliberate attempt to evade plagiarism detection, but I am looking for insights before making any direct accusations.

Has anyone experienced similar issues? Could this be a known student tactic, or possibly a technical glitch with Blackboard/SafeAssign?

Answer 1

The official Blackboard FAQ contains the following text under the Question "What are SafeAssign's supported file types?":

SafeAssign can only process PDFs if the PDF contains machine-readable true text. If the PDF was prepared from "Print to PDF" or scanned from an imaging device without external optical character recognition (OCR), such as a scanner or cell phone app, then the PDF contains a bitmap or vector "picture" of text. This is not readable as text and SafeAssign currently does not perform OCR. If you can't select the text with your mouse or keyboard in a PDF reader, then there is no true text in the PDF.

To prevent this issue: make sure to export the document from your word processor using the "Export to PDF" option and avoid the "PDF Printer" option from the print menu. Also, don't use the option to 'outline' fonts sometimes found in advanced applications because it replaces the text with vector graphics of the same shape which aren't machine-readable.

So, I would emphasize in class broadly, and likely a specific message to the person with the issue, how to make a SafeAssign-readable PDF.

Attributing malice to this seems unwarranted. Now, if the student persists in the behavior then I would insist on them submitting a readable file in some other format listed in the FAQ.

Answer 2

The answer by @Jon Custer pretty much listed out the reason as for why this happens. The TLDR from the FAQ is basically: The PDF are pages of images of texts, instead of pages of texts, and SafeAssign cannot process image based PDFs.

However, I guess where my personal sentiments differ from the consensus here is that I think this is a fair reason to at least be suspicious (without obviously outright accusing the student of cheating). My last contact with Academia was essentially at the dawn of these plagiarism checkers, and back then this was a super common problem.

Now, the reason why I would personally be suspicious is because most (certainly the popular ones like Word) text editors that offers to convert their documents to PDFs, would almost always convert to a PDF that contains true text. After all, there is very little reason to NOT convert a text document to a true-text PDF. The only legitimate reason I can think of is compression, which unless your document is literally thousands of pages long, does not really merit an argument.

However, maintaining the "innocent until proven guilty" motto, I do recommend restricting the ways someone can convert/construct a PDF. For example, only allowing a word-exported PDF or similar, and specifying that invalid PDFs will not be graded, instead of outright reporting this student to an academic misconduct committee.