Hackers can take control of your computer with a single click.

This is exactly what we saw on Tuesday at Google DevFest 2025 Haifa.

In a room packed to capacity with 118 developers, 5 brilliant minds revealed the future of cybersecurity and artificial intelligence:

Stav Cohen (Senior Security Researcher at Zenity) demonstrated how a single malicious prompt can turn AI assistants into weapons.

Or Brokman (Strategic Security Engineer at Google Cloud Security) unveiled the two-sided battle of GenAI. How the same technology serves both attackers and defenders.

Amit LeVi and Rom Himelstein (Research Leads at ATLAS Lab - AI Trust, Learning, Architecture, and Security at Technion - Israel Institute of Technology) and Professor Avi Mendelson, Head of the Lab, took us on a journey from today's LLMs to tomorrow's multi-agent systems, and uncovered security problems that still have no solution.

As a manager in the Google Developers Group Haifa community, I am proud of our amazing community that continues to lead and advance the technological discourse in Haifa and northern Israel.

Thanks to the wonderful speakers for sharing their knowledge and time; we received great feedback from the event.

Thanks to my partners on the community team: Elay Guez, Samer Saida, and Nofar Terenyo.

A huge thank you to our partners at Google Cloud Security and Matam Tech Park for the wonderful hosting, and to everyone who came from across the country to be part of the experience.

Boaz Maoz | Yael Koren | Hagar Hashiloni | Elena Grazi | Alexis Einhorn

Photography: Yalon Schori the champion

My question to you: What is the most frightening or fascinating thing you've heard recently about AI security?

Originally posted on LinkedIn — join the discussion and share your thoughts in the comments.

Hackers can take control of your computer with a single click.

Grow with Roee