AI Detection: What You Need to Know
How Turnitin, GPTZero, and other AI detection tools work—and how to protect your authentic dissertation from false positives.
Important Disclaimer
How AI Detection Works
AI detectors analyze text using multiple signals to distinguish human from machine writing.
Perplexity Score
Measures how 'surprising' the text is. AI text tends to be more predictable with lower perplexity.
Burstiness
Measures variation in sentence length and complexity. Humans write with more natural variation.
Token Probability
Analyzes likelihood of word sequences. AI tends to choose statistically probable words.
Stylometric Analysis
Examines writing style fingerprints like vocabulary richness and syntactic patterns.
Major AI Detection Tools
These are the primary tools universities use to detect AI-generated content.
Turnitin AI Detection
Most universities
Integrated into existing plagiarism workflow
GPTZero
Popular standalone
Free tier available for testing
Originality.AI
Academic & publishing
Strong focus on ChatGPT detection
Copyleaks
Enterprise & education
Multi-language support
The False Positive Problem
AI detection isn't perfect. Here's what can trigger false positives.
Higher Risk of False Positives
- Non-native English speakers
- Highly technical or formulaic writing
- Common academic phrases and structures
- Heavily edited or polished text
- Text on widely-discussed topics
- Short text samples (<250 words)
Lower Risk of False Positives
- Unique personal voice and style
- Domain-specific terminology
- Complex, nuanced arguments
- Personal anecdotes and reflections
- Inconsistent writing quality (human!)
- Longer text samples (1000+ words)
Key Insight
Protecting Your Authentic Work
Strategies to ensure your genuine dissertation isn't falsely flagged.
Document Your Process
Keep drafts, notes, and revision history to prove authentic authorship
- Save multiple dated drafts
- Use track changes or version control
- Keep research notes and outlines
- Screenshot your writing sessions
Maintain Your Voice
Write authentically in your own style, even when using AI for editing
- Don't over-polish to perfection
- Keep discipline-specific terminology
- Include personal reflections where appropriate
- Let your expertise show through
Use AI Transparently
If you use AI tools, disclose them properly in your methodology
- Document which tools you used
- Explain how you used them
- Distinguish AI assistance from your original work
- Follow your institution's disclosure requirements
Know Your Rights
Understand the appeal process if falsely flagged
- Read your university's AI policy
- Know the academic integrity process
- Prepare evidence of authentic work
- Seek support from your supervisor
If You're Falsely Flagged
Steps to take if your authentic work is incorrectly identified as AI-generated.
Don't panic
False positives happen. An initial flag doesn't mean you're guilty of misconduct.
Gather evidence
Collect drafts, notes, browser history, writing session timestamps, and any other proof of your writing process.
Request the full report
Ask to see the complete AI detection report, including confidence scores and flagged sections.
Prepare your explanation
Document your writing process, explain your research methodology, and show how your expertise shaped the content.
Engage with the process
Meet with the academic integrity committee. Be calm, factual, and prepared to demonstrate your authentic authorship.
Know your rights
Understand your institution's appeals process and seek support from your supervisor, student union, or academic support services.
Frequently Asked Questions
Check Your University's AI Policy
Use our AI Policy Checker to understand exactly what your institution requires.