A Plagiarism Checker Flags an Unpublished Manuscript, and the Matches Point in Two Directions
In September, a writer ran an unpublished manuscript, meaning a draft that has not yet been published, through a plagiarism checker. The tool flagged two lines as close to other writing. Publishers Lunch, the publishing industry newsletter, reported the episode on 23 October 2026 in an article by Katy Hershberger.
The checker belongs to Pangram, which calls its offering the Similarity & AI Checker. Plagiarism means presenting someone else's words as your own. Pangram says its plagiarism tool compares a submitted text against webpages, books, news articles and other material to find matches. It added the feature in August 2025.
Neither flagged line was a full copied sentence. Both were common phrases, the kind of wording that turns up in many places without anyone copying anyone. One flag led to a post in an online forum. Another led to the full text of a published book, hosted on Notability in pirated form, meaning copied and shared without the permission of its rights holders.
That second match is the one that makes the story more than a note about a tool. A pirated copy is a different kind of problem from a stock phrase, and the report draws a line between the two. The matches, in other words, did not all mean the same thing, and a flag on its own does not say which kind of match it is.
Pangram's tool is built for people who want to check a text before it goes further, whether that is a student essay, a blog post or a book draft. A flag is a prompt to look closer. It is not a verdict on the writer.
The article's headline also raises questions about where the tool's matches come from. For a writer, that question matters as much as the result itself. A checker is only as useful as the material it compares a draft against, and the public can see only part of that material.
For authors, the practical lesson is simple. Check the flag, read the match, and decide for yourself what it means. For everyone else, it is a reminder that a line of text can be flagged for many reasons, and that the reasons are worth reading before anyone draws a conclusion.
