AI security in 2025 is drastically different from just a few years ago. Organizations have moved beyond experimentation, integrating generative AI into critical business processes. With this shift, AI security has become a priority, requiring a structured, multi-stakeholder approach. While AI security is still evolving, organizations must address traditional security issues alongside AI-specific risks to ensure robust defenses.

In this session, Dr. Anton Chuvakin of Google Cloud shares insights on:

• The evolving challenges of securing AI as it transitions from pilots to production;
• How adversaries are leveraging AI and why their use isn't yet a game changer;
• Emerging best practices for AI governance, emphasizing infrastructure, data, models and applications;
• The promises and risks of agentic AI, particularly in automating security operations;
• The impact of AI on cybersecurity resilience and why short-term gains remain uncertain.