PAN'25 Voight-Kampff Generative AI Detection

Code and data for the Voight-Kampff Generative AI Detection Task at PAN'25 (Subtask 1).

The repository contains the original source code for the baselines and Tira evaluator, as well as code for evaluating the task results.

The training data for the task can be found on Zenodo.

The test data will be kept confidential for future editions of the task.

Installation

Install dependencies via Poetry:

$ pip install poetry
$ poetry install

Cite

You can cite the task as follows:

@InProceedings{bevendorff:2025d,
  author =                   {Janek Bevendorff and Yuxia Wang and Jussi Karlgren and Matti Wiegmann and Maik Fr{\"o}be and Akim Tsivgun and Jinyan Su and Zhuohan Xie and Mervat Abassy and Jonibek Mansurov and Rui Xing and Minh Ngoc Ta and Kareem Ashraf Elozeiri and Tianle Gu and Raj Vardhan Tomar and Jiahui Geng and Ekaterina Artemova and Artem Shelmanov and Nizar Habash and Efstathios Stamatatos and Iryna Gurevych and Preslav Nakov and Martin Potthast and Benno Stein},
  booktitle =                {Working Notes of CLEF 2025 -- Conference and Labs of the Evaluation Forum},
  editor =                   {Guglielmo Faggioli and Nicola Ferro and Paolo Rosso and Damiano Spina},
  month =                    sep,
  publisher =                {CEUR-WS.org},
  series =                   {CEUR Workshop Proceedings},
  site =                     {Madrid, Spain},
  title =                    {{Overview of the ``Voight-Kampff'' Generative AI Authorship Verification Task at PAN and ELOQUENT 2025}},
  year =                     2025
}

You can cite PAN'25 like so:

@InProceedings{bevendorff:2025c,
  address =                  {Berlin Heidelberg New York},
  author =                   {Janek Bevendorff and Daryna Dementieva and Maik Fr{\"o}be and Bela Gipp and Andr{\'e} Greiner-Petter and Jussi Karlgren and Maximilian Mayerl and Preslav Nakov and Alexander Panchenko and Martin Potthast and Artem Shelmanov and Efstathios Stamatatos and Benno Stein and Yuxia Wang and Matti Wiegmann and Eva Zangerle},
  booktitle =                {Experimental IR Meets Multilinguality, Multimodality, and Interaction. Proceedings of the Sixteenth International Conference of the CLEF Association (CLEF 2025)},
  editor =                   {Jorge Carrillo-de-Albornoz and Julio Gonzalo and Laura Plaza and Alba Garc{\'i}a Seco Herrera and Josiane Mothe and Florina Piroi and Paolo Rosso and Damiano Spina and Guglielmo Faggioli and Nicola Ferro},
  month =                    sep,
  publisher =                {Springer},
  series =                   {Lecture Notes in Computer Science},
  site =                     {Madrid, Spain},
  title =                    {{Overview of PAN 2025: Voight-Kampff Generative AI Detection, Multilingual Text Detoxification, Multi-Author Writing Style Analysis, and Generative Plagiarism Detection}},
  year =                     2025
}