We are excited to announce the publication of the talks from the workshop “AI in Cybersecurity: Offensive Capabilities, Impacts, and Response Strategies“. For this workshop, ELSA Coordinator Mario Fritz invited speakers from renowned universities, companies, and the EU AI Office to present their experiences, ideas, and research.
Workshop Topic
AI has boosted capabilities in many domains – or has increased the capabilities of individuals. The cybersecurity domain is no different. AI has been pivotal in automatically detecting cyber incidents in real time with remarkable accuracy, as well as in finding and fixing bugs and vulnerabilities at scale. But it has become increasingly clear that we are also seeing an increasing boost in offensive capabilities. The EU AI office has a dedicated thread to monitor cyber capabilities for General Purpose AI (GPAI) models, and recent research has also demonstrated the emerging capabilities of AI models – e.g., tackling capture-the-flag challenges autonomously. This workshop is dedicated to bringing together researchers, practitioners, industry, and policymakers to understand risks and opportunities in the rapidly changing cybersecurity landscape due to AI. Furthermore, the topic will be strategies and actions that should be in place to do justice to this pressing issues.
Recordings Available on Youtube
Below are the recordings published on YouTube. You can find the abstracts on this workshop website.
Workshop introduction – ELSA Coordinator Professor Dr. Mario Fritz, CISPA Helmholtz Center for Information Security
“Machine Learning for Cybersecurity in Adversarial Settings: Open Issues and Research Challenges” – Prof. Dr. Battista Biggio, University of Cagliari
“ZeroDayClock: Rethinking Defense for an Age of Instant Exploitation” – Sergej Epp, Sysdig
“Red-Teaming the AI Red Team” – Dr. Dario Pasquini, Craken
“From Cyber Warfare to AI Warfare: The Path Ahead” – Professor Dr. Vincent Lenders, University of Luxembourg
“LMs as Penetration Testing Agents: Surprising Capabilities and Ethical Dimensions” – Professor Dr. Jürgen Cito, TU Vienna
“Risk Modelling for AI Safety Evaluations and the Use of MITRE ATT&CK” – Manuel Reinsperger, European Commission AI Office
“Modeling LLM-Enabled Cyber Offense Risk – Current State and Open Questions” – Malcolm Murray, SaferAI
The workshop was part of the ELSA General Assembly 2026.
We want to thank our speakers for sharing their insights not only with the workshop participants but also with the ELSA network and audience.