Copyleaks AI Logic

What is AI Logic?

“AI Logic” is Copyleaks’ enhanced AI content detection module. It doesn’t just flag whether a text might be AI-generated — it seeks to explain how and why a text is flagged.

Key components:

• AI Source Match: This feature checks whether sections of a submission overlap with content already published, including known AI-generated or AI-assisted text. By linking passages back to existing sources, it helps reveal when AI tools may have reused or repurposed material instead of producing something fully original.

• AI Phrases: Copyleaks analyzes linguistic patterns that occur far more frequently in AI-generated writing than in human writing. When these patterns appear in student work, the system highlights them so you can quickly see which passages may have been influenced by AI models.

• Resources: Built into the AI Content area of the Similarity Report, this section includes step-by-step guides, videos, and FAQs to help you understand how Copyleaks detects AI-generated content alongside traditional plagiarism. These resources walk you through the fundamentals, answer common questions, and offer strategies for discussing AI use with your students.

• Transparent “Why”: Instead of simply flagging a passage as AI-generated, Copyleaks shows the reasoning behind the detection. Explanations and signals make the process more transparent, giving you insight into why the system identified a section as likely AI-written.

The goal is to move from “this might be AI” to “this part is suspicious because of X, Y, Z.”

Access AI Content Tab in the Copyleaks Similarity Report

1. Open your Canvas course and go to the assignment you want to review.
   Open SpeedGrader for any student submission you want to inspect.
   

2. In SpeedGrader, locate the colored percentage bubble (the Similarity Score) on the right side of the SpeedGrader toolbar/pane and click it. That opens the full Copyleaks Similarity Report in a new window/tab. 
   

3. In the Copyleaks report, look for the top or side tabs/panels. You should see five tabs. Select the AI Content Tab.
   

4. Once you're inside the AI Content Tab, you will see an overview called AI Content (1), and then two AI Logic sections: AI Source Match (2) and AI Phrases (3). 
   

AI Content Overview

The AI Content section at the top is essentially an overview of AI content found in the student's submission. This percentage combines all of the pieces of CopyLeaks' AI detection. 

1. AI Content Found (Percentage): This is the overall percentage of content from the submitted text that is likely to contain AI Content. 
   

2. Select the triangle to open up a more detailed overall view showing overall percentage of AI vs human text. 
   

AI Logic: AI Source Match

The AI Source Match feature shows you the "why" behind AI detection with sources you can see and verify. AI Logic uses a layered detection approach that applies multiple analysis methods in tandem, providing a highly accurate perspective on how AI may have influenced a piece of text. 

• Navigate to the AI Logic area directly under the AI Content overview. 
  

• Navigate to the AI Source Match AI percentage bar (in this screenshot, it says 50.1%) and hover over it.  
  
  This will open up a detailed panel comparing the student's submission against confirmed AI-generated content already published elsewhere. It shows what percentage of your student's submission is identical to the AI text, similar with only minor changes, paraphrased, and original work. The screenshot below estimates that the student's work is 50.1% paraphrased from a confirmed AI text published elsewhere.
  

• Select Matching Results. 
  

• The Matching Results screen is divided into two panels: the student’s submission appears on the left, and the source it was matched with is displayed on the right. Highlighted passages in the student’s text indicate potential matches. When you click on a highlighted section in the left panel, the corresponding matched text will appear in the right panel for comparison.
  

  Note: Once you select Matching Results, you'll be routed to the Matching tab in the Copyleaks Similarity Report menu, where you've traditionally looked to see if students are plagiarizing published texts. That’s because the AI Source Match feature combines both plagiarism detection and AI detection. Here’s why: plagiarism detection looks for borrowed or stolen work from existing sources, while AI detection looks for writing produced by tools like ChatGPT. But AI tools often generate text by rephrasing or repurposing content they were trained on. In a sense, AI is “plagiarizing with paraphrasing.” That’s why, as you investigate the results in the AI source match area of the report, you'll end up going back and forth between the Matching and AI Content tabs. Copyleaks is showing you not just that a passage may have been AI-written, but also how that AI-generated text connects back to existing source material. 

Recommendation: There is a real possibility that your students will cite sources that appear legitimate, but that Copyleaks flags as written by AI. Students in this scenario will get high AI scores and likely come to you expressing that they didn't know the sources they cited were written ChatGPT etc. Our recommendation is to ensure that guidance on the following is published very prominently in your course and on each assignment:

• What counts as a legitimate source? 

• How to assess if a source is legitimate

• How much (if any) AI use you allow and what the benchmarks are (make sure you've noted it in your Syllabus as well).

This will ensure students have the tools necessary to find and use valid sources not created by AI large langage models. 

AI Logic: AI Phrases

AI Phrases detects phrases that appear with higher frequency in AI written text than in human writing. This does not mean that the phrases are more likely to be written by AI. Rather, it means that they are used more often by AI than by humans. 

1. Navigate to the AI Phrases area. In the top right corner, you'll see how many phrases were flagged (1). Select the See phrases button (2) to view the phrases. 
   
2. In this area, the AI phrases found are divided into three categories based on how much more frequently they are found in AI writing vs human writing: high frequency, medium frequency, and low frequency. In this screen, you can see the frequency ratio (1) and the phrase (2). Select the down caret (3) to view more information about a particular phrase (4). 
   
   
   Note: The AI Phrases feature is not stating that the phrase is only used by AI large language models like ChatGPT. It's comparing how often AI models use the text with how often humans use the text, based on available data. In the above screenshot, the phrase "understanding these differences is essential for" is used 733 times more frequently by AI as opposed to by humans. 

Copyleaks and AI Detection Resources

SPC Guide: Copyleaks FAQ and Troubleshooting

SPC Guide: Getting Started with Copyleaks at SPC

AI Content Detector FAQs

Bringing AI Into the Classroom: Talking to Students about AI

Have Questions? Contact the Academic Technology Support Team

atst@spcollege.edu

 Phone at 727-341-3500

(Hours: Monday - Friday 8:30 a.m. - 4:30 p.m. EST).

After-hours or weekend assistance? Contact the Technical Support Center at 727-791-2795 or onlinehelp@spcollege.edu

Want to Schedule a One-on-One or Group Session on Canvas or other instructional technology issues or opportunities?

Schedule 1:1 Support Help with a Member of the Academic Technology Support Team