What question did this study set out to answer?

The aim is to develop and validate methods for detecting AI-generated texts and understanding their limitations.

April 23, 2026Open Access

Learning to detect AI texts and learning the limits (published version, datasets and other support information)

Key Points

The aim is to develop and validate methods for detecting AI-generated texts and understanding their limitations.
Utilized a comprehensive dataset of AI-generated texts for detection.
Developed detailed lab protocols for reproducibility.
Supported by a project focusing on human-centred AI.
Provided insights into the effectiveness of detection methods.
Identified limitations in current AI text detection capabilities.
Made datasets and protocols publicly available for further research.

Abstract

This repository contains the published article: Milička J, Marklová A, Drobil O, Pospíšilová E (2025) Learning to detect AI texts and learning the limits. PLoS One 20 (10): e0333007. https: //doi. org/10. 1371/journal. pone. 0333007 And all datasets needed to reproduce the research, including a lab protocol that describes in detail all steps necessary. This work was supported by the project “Human-centred AI for a Sustainable and Adaptive Society” (reg. no.: Z. 02. 01. 01/00/23₀25/0008691), co-funded by the European Union (specifically the development of the experimental software)

Read Full Paperexternally

Bookmark

View Full Paper

Cite This Study

Milička et al. (Sat,) studied this question.

synapsesocial.com/papers/69e9ba2a85696592c86ec6c2 https://doi.org/https://doi.org/10.5281/zenodo.19686678

Bookmark

View Full Paper