does Turnitin actually detect QuillBot? tested it

Everyone wants to know if Turnitin can detect QuillBot. I ran some tests to find out what actually happens when you submit QuillBot-paraphrased text.

The Test Setup

I took three scenarios and ran each through Turnitin:

1. Human-written text paraphrased through QuillBot
2. ChatGPT text paraphrased through QuillBot
3. ChatGPT text heavily paraphrased through QuillBot (multiple passes)

Results

Scenario 1: Human text + QuillBot
AI detection score: 5-8%. Turnitin barely flagged it. Makes sense because the underlying content is human-written, QuillBot just adjusted the wording. Plagiarism score was also low since the text was original.

Scenario 2: AI text + QuillBot (single pass)
AI detection score: 35-55%. QuillBot reduced the AI score significantly from the original 95%+ but didn’t eliminate the AI signal. Turnitin still caught the underlying AI patterns.

Scenario 3: AI text + QuillBot (multiple passes)
AI detection score: 15-25%. Running text through QuillBot multiple times with different settings reduced the AI score further but introduced awkward phrasing. The text read unnaturally, which might raise separate red flags with a professor.

What This Tells Us

Turnitin doesn’t specifically detect “QuillBot.” It detects patterns associated with AI-generated text. QuillBot changes the surface-level wording but doesn’t fundamentally alter the deep statistical patterns that AI detectors look for. It helps, but it’s not a reliable bypass.

The Bottom Line

Using QuillBot on your own writing? You’re fine, Turnitin won’t flag it. Using QuillBot to disguise AI text? It reduces the score but doesn’t eliminate detection reliably. For that use case you’d need a purpose-built AI humanizer, and even then the ethical question remains.