Why Turnitin in Canvas Sometimes Shows No AI Score at All

You opened your Turnitin report in Canvas, found your similarity percentage right where it always sits, and then noticed the AI writing indicator is just… empty. No number, no flag, no explanation. Before you assume something broke or start reading meaning into the blank, here’s the honest version: a missing AI score is almost always expected behavior, and it rarely says anything about your writing.

Key takeaways

• A missing AI score in Canvas is usually expected behavior, not a bug, and it doesn’t mean your paper passed or failed anything.
• The single most common cause is length. Turnitin requires at least 300 words of prose before it will generate an AI indicator.
• Instructors and institutions can turn the AI writing indicator off entirely, so the column can be blank for an entire class.
• File type, language, timing, and the gap between the student view and the instructor view all change whether a score appears.
• The AI indicator is a probability estimate, not proof, so its absence or presence shouldn’t be read as a verdict.

First, what the AI indicator actually is

Turnitin’s AI writing detection is separate from its similarity check. Similarity compares your text against a database of papers, journals, and web pages to find overlap. The AI indicator does something different: it estimates the percentage of the submission that reads like it was generated by a large language model, based on statistical patterns in the writing rather than matches to any source.

That distinction matters here, because the two run as independent processes. You can have a perfectly normal similarity report and a completely blank AI indicator in the same Canvas submission, and nothing is wrong. If you want the deeper mechanics of how these tools score text, we’ve written about the statistics behind how detectors decide. For now, hold onto one idea: the AI score is a probability, and Turnitin only produces one when it has enough text to work with.

The most common reason: your paper was too short

This is the big one. Turnitin’s AI writing detection has a minimum-length requirement, and it’s stricter than most people expect. Turnitin raised that floor from 150 to 300 words, and the submission now needs at least 300 words of prose before the indicator runs. Fall under that and you get no score at all, no matter how the text was written. Turnitin’s own reasoning is that accuracy improves with more text, so the tool focuses on long-form writing.

And “300 words” is sneakier than the word count at the bottom of your document. Turnitin counts qualifying prose only, and strips a lot out first:

• Quotations — directly quoted passages don’t feed the AI analysis.
• Bibliography and reference lists — citations are excluded.
• Bulleted or list-heavy content — fragments and short lines aren’t long-form prose.
• Non-English text — the indicator analyzes English-language prose; other languages are skipped.

A quick scenario

Maya submits a 480-word reflection for her Canvas-linked assignment. Her document word count looks safe. But about 140 words of it are block quotes from the assigned reading, and another 90 are a formatted reference list. Turnitin strips both, leaving roughly 250 words of analyzable prose. That’s under the threshold, so the AI indicator never generates. Maya sees her similarity score and an empty AI field, panics for a second, then realizes her paper simply didn’t have enough qualifying text for the tool to bother.

If your assignment is short, quote-heavy, or mostly lists, a blank AI score is the expected outcome.

The feature might be turned off entirely

Here’s something a lot of students, and even some instructors, don’t realize: the AI writing indicator is controlled at the institution and account level. A Turnitin administrator can disable it, and individual instructors may have it switched off for a given assignment.

When that’s the case, the AI column is blank for *everyone* in the class. There’s no score to show because the analysis was never enabled. This has become more common rather than less, as some universities have stepped back from leaning on AI detection. We’ve covered the broader trend of schools rethinking these mandates over on the blog.

So if you’re comparing notes with classmates and nobody has an AI score, that’s a strong sign the feature is simply off for your course. It’s not a glitch, and it’s not something you triggered.

Student view versus instructor view

This one trips people up constantly. By default, the AI writing indicator is visible only to instructors and administrators inside the Turnitin viewer. As a student looking at your own submission in Canvas, you usually see your similarity report and nothing else.

So when a student says “Turnitin isn’t showing my AI score,” the honest answer is often that it was never meant to show it to you. The number may exist on the instructor’s side while your view stays blank by design. If you’re a student worried about what an instructor might see, that’s a different question from whether the tool malfunctioned, and it’s worth a calm conversation rather than guesswork.

File type and formatting problems

Turnitin has to read your submission as text. If it can’t, the AI indicator won’t run even when the paper is long enough.

The usual culprits:

• Scanned PDFs or images of text. If you photographed or scanned handwritten or printed pages, Turnitin sees pixels, not words. No readable prose, no AI analysis.
• Unsupported or unusual file types. Stick to the formats Turnitin lists for AI reports — .docx, text-based .pdf, .rtf, and .txt. Exotic formats may upload but fail to process for AI detection.
• Corrupted or password-protected files. If the file won’t open cleanly on Turnitin’s end, neither report may generate properly.

Another quick scenario

Diego writes a solid 1,200-word essay, exports it as a PDF, but the export turns each page into an image rather than selectable text. His similarity score comes back oddly low and his AI indicator is blank. The reason is the same for both: Turnitin couldn’t extract any actual text. The fix is to resubmit a text-based document, not to assume the detector cleared him.

Timing and processing delays

The similarity report and the AI indicator are produced on separate passes, and they don’t always finish at the same moment. Right after you submit, especially near a deadline when thousands of students are uploading at once, you may see the similarity score populate first while the AI field sits empty.

That’s not a final state. Refresh after a few minutes. During peak load, generation can lag noticeably. A blank score immediately after submission tells you almost nothing, so wait before drawing conclusions.

When none of these explain it

Say your paper is well over 300 words of clean English prose, the feature is enabled, you’re viewing as an instructor, the file is a standard text document, and you’ve given it time. If the AI score still won’t appear, then you’re probably looking at a genuine technical issue. At that point it’s a Turnitin support or institutional-admin question, not something you can fix from the Canvas side. Instructors should loop in their Turnitin administrator rather than re-uploading the same file over and over.

Don’t read meaning into the blank

This is the part worth slowing down on. A missing AI score is not a passing grade, and it’s not a failing one. It’s the absence of a measurement. Turnitin’s own guidance describes the AI indicator as a percentage estimate, not proof of misconduct, and the tool has real limits — short text breaks it, and detectors in general produce false positives. A 2023 study by Stanford-affiliated researchers (Liang et al., published in *Patterns*) found that GPT detectors disproportionately flag writing by non-native English speakers. OpenAI even retired its own AI classifier in July 2023, citing a low rate of accuracy. These tools output probabilities under specific conditions, and one of those conditions is having enough qualifying text to analyze in the first place.

So treat a blank the way you’d treat a thermometer that never got placed in the oven: it didn’t run, so there’s no reading. The smart move isn’t to reverse-engineer the detector. It’s to write clearly in your own voice, cite honestly, and make sure your submission is long enough and clean enough that the work speaks for itself. If you ever want to understand how your own draft reads to these systems before you turn it in, that’s a reasonable, white-hat thing to check, and how we frame fair use is laid out on our plans page.

Frequently asked questions

The Q&A below covers the situations students and instructors run into most, from the student-versus-instructor view to processing delays.

The bottom line

Nine times out of ten, a missing Turnitin AI score in Canvas comes down to length, settings, file type, view permissions, or timing — not a broken tool and not a hidden verdict. Understanding why the field is blank beats panicking over it. If you want to see how your own writing reads before you submit, you can check a draft yourself and decide what to do with the result on your own terms.