Jan Franciszek Piotrowski

“Try to leave this world a little better than you found it,” said Robert Baden-Powell. Based on this philosophy, I would like to use my talents to make AI beneficial and safe, and to reduce the risks associated with it.

I started my ML journey as part of a project on Twitter disinformation, which led to publications at EMNLP 2024. During this time, I worked commercially and also tried interdisciplinary projects, ultimately deciding on AI Safety. My goal is to do research that informs public policy - evaluations, data, and recommendations that decision-makers can use.

Privately, I assume that you guessed after the first quote, I am a scout leader. And I will be honest - I gain an incredible amount of energy from engaging people in good things.

selected publications

LLM generated responses to mitigate the impact of hate speech

Jakub Podolak, Szymon Łukasik, Paweł Balawender, and 4 more authors

In Findings of the Association for Computational Linguistics: EMNLP 2024, Nov 2024

Abs DOI HTML

In this study, we explore the use of Large Language Models (LLMs) to counteract hate speech. We conducted the first real-life A/B test assessing the effectiveness of LLM-generated counter-speech. During the experiment, we posted 753 automatically generated responses aimed at reducing user engagement under tweets that contained hate speech toward Ukrainian refugees in Poland.Our work shows that interventions with LLM-generated responses significantly decrease user engagement, particularly for original tweets with at least ten views, reducing it by over 20%. This paper outlines the design of our automatic moderation system, proposes a simple metric for measuring user engagement and details the methodology of conducting such an experiment. We discuss the ethical considerations and challenges in deploying generative AI for discourse moderation.
[ENG - Translated] Stanisław Lem Predicted the Pathologies of X — Can Elon Musk’s Platform Be Saved?

Jan Franciszek Piotrowski

Nov 2025

Klub Jagielloński, 23 Apr 2025

HTML