The Polish Association for National Security is implementing the project „R-GRID: AI Algorithms for Threat Prediction and Power Grid Protection”. It is an innovative simulator for resilience assessment and crisis exercise planning, enabling operators and public institutions to plan and prepare for potential crises more effectively. The goal is to create a tool that not only helps better understand possible crises, but also allows users to test solutions to those crises.
On Monday (8th December 2025), as part of this project, workshops for young researchers were held at Rzeszów University of Technology. Among the topics discussed in the individual sessions were: „Hybrid threats to Poland’s critical energy infrastructure”, „The use of artificial intelligence to enhance the state’s resilience in the energy sector”, and „Scenarios of future energy crises”.
How to Deal with Malicious AI
Dr. Klaudia Rosińska, an expert at the Polish Association for National Security, spoke about the threats associated with the use of „malicious AI”. She used this term to describe entire language models developed to negatively influence people, as well as malicious overlays that infect models created in Western countries. In her view, „malicious AI” enables profiling for cyber aggression, manipulation, and persuasion.
„Since troll farms have been created, we can assume that analogous farms already exist where language models are being trained. These trainers prompt the algorithms in such a way as to increase their effectiveness and precision in influence operations and malicious activities”, the expert notes.
Dr. Rosińska presented scenarios for the use of „malicious AI”. The first was deep influence operations targeting individuals rather than, as before, large social groups. A chatbot is capable of tailoring its arguments to the personality profile of the person it interacts with.
„There is a need to create a monitoring system in Poland for Russian propaganda entities”, argues Klaudia Rosińska. „It is also important to develop systems that identify content, as well as systems that mark trustworthy content (‘watermarks,’ government certifications), especially in the energy sector. The goal is to distribute truthful content while removing false information”.
Another dangerous use of chatbots may involve automated recruitment on behalf of hostile intelligence services. Democratic states should therefore invest in research into such recruitment techniques and work to eliminate them. Moreover, an appropriate legal framework is necessary – one that, so far, does not exist.
Vulnerable Areas
Mr Maciej Kluczyński, President of the Polish Association for National Security, notes that an adversary conducting hybrid operations against Poland is particularly interested in four critical infrastructure systems essential for maintaining the continuity of state functions and public structures. These are energy, the transport of people and goods, communications, and water supply systems. Examples include regular cyberattacks on water and sewage system software. It can be assumed that, as part of hybrid warfare, Russia will aim to trigger small ecological disasters.
The expert also mentioned observed operational patterns of Russian services, particularly in the area of energy security. These include obstructing investments aimed at diversifying supplies of electricity or fuels; disrupting the operation of critical energy infrastructure facilities; and sabotaging planning and implementation processes of such investments. Finally, this also includes identifying vulnerabilities in the security systems of critical energy infrastructure facilities and enabling disruptions to their operational continuity.
The number of incidents involving security breaches of critical infrastructure facilities in Poland is increasing. Between 2022 and 2025 alone, 20 related criminal cases have been concluded, and 40 additional cases are ongoing.
Who Is Responsible for Security?
Prof. Tomasz Michalak from the University of Warsaw, researcher at the IDEAS Research Institute, emphasized in his presentation that absolute security is never fully attainable. In his view, a reasonable approach is randomization, meaning the selection of the most probable directions from which threats may emerge. Professor Michalak also discussed specific examples of solutions used to ensure security, illustrated with the case of Los Angeles International Airport.
Speaking in one of the panels, Prof. Mariusz Ruszel from the Rzeszów University of Technology and President of the Łukasiewicz Institute for Energy Policy, stressed that the boundary of responsibility between the state and an energy company in the area of energy security is defined by public–private partnership.
In addition to the experts and researchers mentioned above, the conference also featured contributions from: Dr. Eng. Krzysztof Łowczowski (PTBN, Poznań University of Technology), Dr. Stanisław Mikulski (Ideas NCBR, Poznań University of Technology), Dominik P. Jankowski (NATO Headquarters), Dr. Eng. Beata Piotrowska (Rzeszów University of Technology, Institute for Energy Policy), and Zuzanna Nowak (The Opportunity Institute for Foreign Affairs, Institute for Energy Policy).
The project “R-GRID: AI Algorithms for Threat Prediction and Power Grid Protection” is funded by the NATO Science for Peace and Security Programme. The project consortium includes the Polish Association for National Security, Ideas NCBR Sp. z o.o., Laurea University of Applied Sciences (Finland), and the Ukrainian Institute for the Future.
