From Session Hijack to Data Leak: Breaking an AI Agent's Trust Boundary
AI agents are increasingly becoming part of applications that handle sensitive data and perform actions on behalf of authenticated users. But what happens when an attacker breaks the trust boundary of an AI agent?
This session delivers a live attack-chain demonstration that starts with session hijacking and ends with unauthorized cross-tenant data access through an AI agent.
-
Edison Plaku
Cybersecurity Analyst
at
National Agency for Information Society (AKSHI), Albania
Enterprise Security Validation at Scale: Building an AI Offensive Security Platform
Manual penetration testing remains essential, but how do you scale it in an era when attackers use AI to penetrate your environment? How do you respond by building solutions that improve testing safely and decide when human expertise should lead and when AI agents can augment the work?
Join us for a practical exploration of AI augmentation in offensive security. We’ll share real-world insights on balancing human judgment, agent capabilities, safety controls, and the resources needed to scale enterprise security validation.
-
Marcos Lustosa
AI & Automation Lead
at
Stefanini
-
Daniel Vaida
AI & Automation Lead for Cybersecurity
at
Stefanini
Attackers know who to target. Do you?
A familiar voice on the phone. A convincing video call. A message that arrives at exactly the right moment. AI has made it easier than ever to exploit the very things we rely on to establish trust.
-
Ovidiu Popeti
Moderator
at
Cybersecurity Con
Your Encryption Has an Expiry Date: Surviving Q-Day and AI-Speed Attacks
Two clocks are running at once. Q-Day is when quantum computers break today's encryption, and under the EU's roadmap, high-risk use cases come off quantum-vulnerable crypto by 2030.
-
Tudor Damian
Cybersecurity, AI & Cloud Advisor
at
D3 Cyber
Translating AI cyber risks to business. Mission (almost) impossible
Translating your classical cyber risk to business or boards was already difficult. With AI added in the mix, mission has become almost impossible. How can security teams fix the basics while keeping pace with AI adoption?
-
Tiberiu Anghel
Group Chief Information Security Officer
at
VEO Worldwide Services
The API Made Me Do It - Do Bad APIs Lead AI to Generate Vulnerable Code?
GenAI code generation is often evaluated as a model problem: the model failed to reason about security, misunderstood the prompt, or reproduced vulnerable examples from its training data.
-
Yariv Tal
Secure Coding Mentor
at
Secure From Scratch
When a Country Goes Dark: The Human Side of Incident Response
A hostile state wiped a national government’s datacentre and I was called in to coordinate the response. This is the story of what happened next: not the technical side, but the people, the politics, the fear and the decisions made under real pressure, with no playbook and no time.
-
Alexandru Suditu
Co-founder & Managing Director
at
SentryOT
"Mihai or the AI”? When the Breach Hits, Who Do You Call?
-
Alexandru Gheorghe
Founder
at
Data Trail