Back to feed

How to Detect AI Plagiarism in 2026

CopyLeaksAI detectorplagiarism checkerCanvasacademic integrityAI-generated contentfalse positiveUndetectable.aistudent submissionoriginality report
How to Detect AI Plagiarism in 2026

Key points

CopyLeaks replaces Unicheck in Canvas by January 2024, adding AI detection. Use it as a tool, but verify findings with traditional methods to ensure accuracy and support academic integrity.

Key takeaway

The transition to CopyLeaks as the sole plagiarism detection tool in Canvas by January 2024 introduces an integrated AI detector, enhancing academic integrity efforts. However, AI detection tools are not infallible; they can be circumvented by services like Undetectable.ai, and high AI scores alone should not justify punitive actions without further investigation. Educators must combine these tools with traditional methods—such as comparing student work, conducting discussions, and reviewing sources—to accurately assess originality. The administration emphasizes that actions based solely on AI detection will not be supported in appeals, underscoring the need for a balanced, evidence-based approach to uphold academic standards while fostering student learning and integrity.

All right, hi everyone. We've got a new plagiarism checker in Canvas called CopyLeaks. You may have noticed it already. Unicheck is still there, but we're keeping it only through the end of December. Starting in January 2024, CopyLeaks will be the only plagiarism checker available. We're giving you time to transition. If you don't want to deal with it now, you can continue through fall, but be aware that in January, there will only be one.

The key reason for switching is that CopyLeaks includes an AI detector—a tool to identify artificial intelligence-generated content. It's available for use, but it's not the ultimate solution. We'll discuss that more. Let's walk through setting up an assignment in Canvas to show how it works. If you've used Unicheck before, the process is similar and straightforward.

First, we'll create an assignment. Click on "Assignments," then "Add Assignment." For example, name it "Inspection and Fire Services." Set the submission type to online, choosing text entry or file upload. You'll see the plagiarism review box appear. Click it and select CopyLeaks. The scan settings are pre-configured, so you don't need to adjust them. The main decision is when students see the report. We recommend immediately, so they can learn about academic integrity, but you can set it to after grading, after the due date, or never.

Once saved and published, the assignment is ready. Now, let's submit a plagiarized example as a student. Imagine a student named Zippy copies and pastes content from an online source. They submit it, and within minutes, a plagiarism report generates. A red flag appears in the grades section. Clicking it shows a 100% match with a source. Always verify by checking the source link to confirm plagiarism. In this case, it's a direct copy, so it's clear. This is when you follow your academic integrity policy, whether that's a zero, a warning, or a chance to resubmit.

Next, let's submit an AI-generated assignment. Zippy uses ChatGPT to create an answer and submits it. The report may show false positives from previous submissions, which can be excluded. After exclusions, the plagiarism score might drop to 0%, but the AI detector flags it with a "suspected cheating AI text detected" alert. It highlights sections likely AI-generated.

However, AI detectors are notoriously inaccurate. They are tools, not definitive proof. According to guidance from the deans and VPI, a high AI score alone is not grounds for a failing grade. You must investigate further—talk to the student, compare work, or generate your own AI response for comparison. Actions based solely on AI scores will not be supported in appeals. So, use the detector as a flag, but rely on traditional methods for confirmation.

To demonstrate why, let's show how AI detection can be bypassed. Zippy uses a service like Undetectable.ai, which modifies AI content to evade detection. After processing, the submission shows 0% plagiarism and no AI flags. This highlights that students can easily access tools to beat AI detectors, so don't rely on them exclusively.

Finally, for submissions outside Canvas—like emailed documents or paper assignments—you can use CopyLeaks directly. Go to apps.copyleaks.com, upload the file or paste text, and scan. It generates a report you can use for discussions with students.

In summary, CopyLeaks is our only plagiarism checker from January 2024, with built-in AI detection. Use it as part of a broader approach, combining tools with investigation and dialogue to maintain academic integrity effectively.

Frequently Asked Questions

Qany questions?

Please read the article carefully. If you have any questions, please contact [email protected].

Audio synthesized by Entity-Echo AI Agent

Playback speedDownload audio