Related Experiment Video
Updated: May 12, 2026

08:58
Artificial Intelligence Approaches to Assessing Primary Cilia
Published on: May 1, 2021
3.4K
How Sensitive Are the Free AI-detector Tools in Detecting AI-generated Texts? A Comparison of Popular AI-detector
Sujita Kumar Kar1, Teena Bansal1, Sumit Modi1
1Dept. of Psychiatry, King George's Medical University, Lucknow, Uttar Pradesh, India.
Indian Journal of Psychological Medicine
|November 20, 2024
Summary
Free AI detection tools vary in accuracy. Some tools detect AI-generated text perfectly, while others struggle with paraphrased content, highlighting the evolving landscape of artificial intelligence in academic writing.
Area of Science:
- Academic Writing
- Artificial Intelligence
- Digital Humanities
Background:
- Artificial intelligence (AI) significantly impacts academic writing.
- Concerns exist regarding the authenticity and originality of AI-generated academic content.
- AI-detection software has emerged to identify AI-assisted or AI-generated text.
Purpose of the Study:
- To evaluate the sensitivity of free AI-detection software in identifying AI-generated text.
- To assess the performance of these tools on both original and paraphrased AI-generated content.
Main Methods:
- Searched for and selected 10 free AI-detection tools.
- Generated a scientific article using ChatGPT 3.5.
- Rephrased the AI-generated text using three different software tools.
- Analyzed original and paraphrased texts with the selected AI-detection software.
Main Results:
- Sensitivity of 10 AI-detection tools ranged from 0% to 100%.
- Five tools achieved 100% accuracy in detecting original AI-generated text.
- Sapling and Undetectable AI accurately identified all paraphrased AI content.
- Copyleaks, QuillBot, and Wordtune detected 100% of paraphrased content from two specific rephrasing tools.
Conclusions:
- Current free AI-detection platforms show varying sensitivity.
- Some tools demonstrate high accuracy for original AI text, but performance varies with paraphrased content.
- Future software updates may improve the detection capabilities of AI-generated academic writing.

