The study by William H. Walters, "The Effectiveness of Software Designed to Detect AI-Generated Writing: A Comparison of 16 AI Text Detectors," provides a comprehensive evaluation of 16 publicly available AI text detectors. Based on the study’s findings, Originality.ai showed remarkable accuracy in identifying AI-generated text.
This study evaluated the accuracy of 16 publicly available AI text detectors in discerning AI-generated from human-generated writing. The evaluated document included 126 essays generated by AI and humans across various domains.
Each detector’s performance was assessed for its overall accuracy, its accuracy with each type of document, its decisiveness, the number of false positives, and the number of false negatives.
Originality.ai was one of the top three highest-performing AI detectors — demonstrating high accuracy with all three sets of documents.
GPT 3.5 and GPT 4 were each used to generate 42 short papers (literature reviews), similar to the type of papers typically assigned to students in first-year composition courses at US universities. The topics covered social sciences, natural sciences, and the humanities.
The 42 human-written documents were gathered from a set of 178 papers submitted by Manhattan College English 110 (First Year Composition) students during the 2014 to 2015 academic year.
The 2014 to 2015 timeframe was selected because it occurred before the widespread availability of AI tools such as ChatGPT to ensure the papers were written without AI.
This dataset consists of 126 documents divided into the following:
Each detector’s performance was assessed for:
The analysis involved four steps:
Originality.ai excelled at detecting both AI-generated and human-written documents with high accuracy. The study positioned it as a top-tier AI-generated text detection tool with a 97% overall accuracy rate.
Originality.ai is positioned as a top-tier AI-generated text detection tool with exceptional accuracy. Its performance is on par with, and sometimes even better than, other leading tools like CopyLeaks and TurnItIn, ensuring the authenticity and integrity of the written content.
Have you seen a thought leadership LinkedIn post and wondered if it was AI-generated or human-written? In this study, we looked at the impact of ChatGPT and generative AI tools on the volume of AI content that is being published on LinkedIn. These are our findings.
We believe that it is crucial for AI content detectors reported accuracy to be open, transparent, and accountable. The reality is, each person seeking AI-detection services deserves to know which detector is the most accurate for their specific use case.