logos

Live SEMINAR UNESCO Chair on the Ethical Governance of Artificial Intelligence - Prof. Maciej Ogrodniczuk

Title: PLLuM – Polish Large Language Model

Broadcast date: 2026-02-26 from 11:00

Link to the broadcast: [LINK] https://www.youtube.com/watch?v=vv042Fcy8wc

Abstract

The development and pilot implementation of Polish Large Language Models (PLLuM) is a strategic initiative of the national public administration. Led by a consortium of leading Polish research institutions, the project aims to create a sovereign AI ecosystem that is tailored to the specific linguistic and legal requirements of the Polish state. This talk will present the project's methodology, results and evaluation, as well as the adaptation of the model for administrative tasks. It will also focus on the use of extensive language resources accumulated by the IPI PAN over many years for the development of the collection of organic instructions, and for training and evaluating the models.

Prof. Maciej Ogrodniczuk
The Linguistic Engineering (LE) Group
Department of Artificial Intelligence
Institute of Computer Science (IPI PAN)
Polish Academy of Sciences
poster
Live SEMINAR UNESCO Chair on the Ethical Governance of Artificial Intelligence - Prof. Włodzisław Duch

Live SEMINAR UNESCO Chair on the Ethical Governance of Artificial Intelligence - Prof. Włodzisław Duch

Title: Ai and the Future of Science

Broadcast date: 2025-12-15 from 11:00

Link to the broadcast: [LINK] https://www.youtube.com/watch?v=k0p4BJ0XooY

Abstract

In September 2025 frontier AI models have surpassed humans in several programming and mathematical competitions. In life sciences systems like AlphaFold, AlphaFold, AlphaGenome, AlphaProteo, AlphaMissense and a dozen others became indispensable. Physics-informed machine learning systems trained to respect constraints based on the natural laws are successfully applied to analyze complex real-world phenomena, from quantum systems to weather prediction. In many branches of science great progress has already been made at identification of new phenomena in experimental data. AI algorithms are also used for design of new experiments, making discoveries at the conceptual level. Superhuman performance is achieved by self-improving models simulating various types of human intelligence. The number of papers written to a significant degree with support of AI systems is already larger than those written by humans. To address grand challenges new platforms supporting scientific research are being created, involving humans supported by teams of specialized AI agents with specific types of intelligence adjusted to the particular needs and domains of research.

poster