My Canvas Assignment Got Flagged for AI — Here's the Detector It's Actually Using

Priya had been back in school for two semesters when it happened. A 34-year-old nursing student at a community college in Florida, she was working double shifts at a care facility while finishing her prerequisites. One Sunday night she submitted a patient care plan through Canvas. By Tuesday morning, her professor had sent a message: the assignment had been flagged for AI writing.

She had used ChatGPT. Just to get the structure started — a skeleton she then rewrote in her own words. Or so she thought.

What AI Detector Does Canvas Actually Use?

Canvas itself does not have a built-in AI detector. Canvas is a learning management system — it is the platform where you submit work, not the tool that scans it. The AI detection happens through third-party integrations, and the most widely used is Turnitin.

When an instructor enables Turnitin inside Canvas, submissions automatically run through Turnitin's AI writing detection feature, which launched in 2023. Some institutions use other LTI-integrated tools like Unicheck or iThenticate, but Turnitin is by far the most common setup. If you're seeing an AI probability score inside Canvas, you are almost certainly looking at Turnitin's output — not a Canvas-native tool.

Understanding how AI detectors work is the first step toward understanding why this happens even to people who genuinely rewrote the content.

Why Priya's Rewritten Draft Still Got Flagged

Priya's situation is not unusual. Turnitin's AI detector doesn't just check your final words. It analyzes writing patterns — sentence rhythm, predictability of word choice, structural uniformity across paragraphs. When someone uses ChatGPT to draft and then edits lightly, the detector often still reads the original AI pattern underneath.

She had rewritten maybe 60% of the sentences. But she had kept the overall flow, the paragraph transitions, the clinical formatting. That was enough to preserve a high AI probability score.

This is a well-documented problem. The issue of AI detection false positives hits hardest when someone has partially edited AI text rather than either fully using it or writing entirely from scratch — the detector sees the ghost of the original structure.

What Priya Did Over the Next Three Days

She had three days before a formal academic integrity review was scheduled. Here's exactly what she did:

• Day 1: She ran her original submission through WriteMask's free AI detector to see where the AI signals were strongest. The paragraph-level breakdown showed her precisely which sections were pulling the score up.
• Day 2: She used WriteMask to restructure the flagged sections — not to obscure wrongdoing, but to replace AI-pattern phrasing with language that matched her actual clinical voice. She reviewed every suggested change against her nursing knowledge before accepting it.
• Day 3: She resubmitted a revised version and emailed her professor proactively, explaining her drafting process and attaching her original handwritten notes as evidence of her thinking.

The revised version cleared the detector. WriteMask has a 93% pass rate on Turnitin's AI detection, and Priya's rewritten care plan was no exception. Her professor accepted the resubmission without further action.

The Bigger Lesson About Canvas and AI Detection

A lot of students treat Canvas as the system doing the checking. It's not. Canvas is just the front door. The actual detection happens in whichever tool your institution has plugged into the LMS — and that varies by school.

Before you submit anything that AI touched, it's worth knowing which detector your campus uses. Many syllabi list this, but if yours doesn't, you can look up your institution's specific Canvas integration at our university AI policies page, which tracks LMS-linked detection tools by school.

What to Do If You're Already Flagged Through Canvas

Don't panic. Don't delete anything. The steps that actually help:

• Request a copy of your Turnitin report — you're entitled to see the breakdown
• Identify which sections scored highest for AI probability, not just the overall score
• Rewrite those sections from scratch using your own sentence rhythms and vocabulary
• Keep every draft, note, and browser history as evidence of your process
• Read up on what to do if you're accused of using AI before any formal meeting — knowing your rights matters

Priya's case wasn't about dishonesty. It was about a tool that created ambiguity. The Canvas plus Turnitin pipeline is powerful, but it is not perfect. Knowing exactly how it works — and which layer is actually doing the detecting — gives you real options when something goes wrong.