Скачать или смотреть Panel: From Risk to Resilience , Securing Agentic Frontiers

🧭🛡️ Securing Agentic Frontiers: Why Agentic AI Needs Its Own Security Playbook
This panel session from the OWASP GenAI Security Project Virtual Summit (October 2025) brings together John Sautopoulos (Agentic Security Initiative co-lead), Ron del Rosario (VP & Head of AI Security, SAP; co-lead), and Chris Hughes (CEO/Co-founder, Aquia) to explain what changes when GenAI moves from chatbots to tool-using agents.

The core message: with agency comes risk. Agentic systems can take actions, chain tools, and operate at massive scale—creating new attack surfaces around identity/privilege abuse, tool & protocol risk (including malicious MCP servers), cascading failures across multi-agent workflows, and supply chain problems. The panel emphasizes that the biggest near-term challenge for security leaders is visibility, understanding, and traceability: knowing where agents exist, what data/tools they can access, and how they reached a decision—so humans can intervene when risk signals trigger.

They also explore how developer tooling (IDE “ask mode” vs “agent mode”) changes risk, why cognitive overreliance and iteration can degrade code security, and how threat modeling + architecture-driven defense-in-depth helps teams move from fear to systematic controls. The discussion closes with why OWASP is building an Agentic AI Top 10—because many “theoretical” threats are already becoming real incidents.

👉 Learn more about the OWASP GenAI Security Project:
https://genai.owasp.org

#AgenticAI #AISecurity #OWASP #GenAI #MCP #ThreatModeling #ZeroTrust #SupplyChainSecurity #AI Governance