Marc Cortesi, Chief Information Security Officer & Timm Süss, Lead AI Circle at Baloise

AI – what the regulation requests and how we interpret it
With the AI Act companies developing and providing AI systems to the EU market have to comply with certain requirements. Baloise tells you how this Act is being translated in the company and how we they deliver the topic of AI to the Users

Rainer Kessler, Chief Scientist & Senior Lecturer (multiple institutions: UZH, UB, FHNW, FHGR, etc. )  

Jailbreaking GenAI: usig good for evil
Dramatic arms-race behind the scenes: A.I. can solve problems for the good of humanity, or it can be used as a digital-weapon of mass-destruction. BigTech is trying to prevent the latter, but circumventing the precautions is possible!

Peter Bury, Domain Consultant at Palo Alto Networks

AI and The CyberSecurity Frontier
AI is a rapidly evolving domain simultaneously presenting benefits and risks to organisations embarking on AI transformation. This talk looks at the significance of AI in cybersecurity, the risks associated with it, the threats from it and the way it can be employed to drive more efficient cybersecurity and operational resilience.

Nicolò Fornari, Penetration Tester at Compass Security

Deep Learning Models: Security Applications
The past year has witnessed an explosion of development and interest in deep learning models, especially with respect to LLMs.
 
This led to the question: how can these models be used in security, either offensive or defensive? The talk will explore some use case, focusing on the process from start to end, in terms of hardware requirements, limitations and results.

The main use case will be on voice cloning. While deep fakes of actors and politicians are well known, they are also "easy wins" given that training material is abundant and english is often the language of choice. The talk will explore the effort and the results to set up a voice clone for any person.