Heading 1

Analyst Briefing with Anton Chuvakin, formerly an analyst at Gartner and now under Google Cloud Security, explores how Straiker secures Agentic AI Applications.

In today's rapidly evolving digital landscape, AI is not just an innovation; it's becoming the core driver of business transformation. There's increasing discourse around securing AI and, specifically, agentic AI.

AI security has to be a priority because AI is going to be the number one threat vector.

Agentic AI introduces a paradigm shift where reasoning models replace traditional business logic layers. As a result, companies can envision applications where human tasks are executed autonomously by these models, given access to necessary tools.

The adoption of agentic AI apps demand a fresh perspective on security. Traditional vulnerabilities such as SQL injections may not vanish overnight but are overshadowed by new threats like prompt injections. To combat this, businesses must consider runtime guardrails for AI apps and AI agents, ensuring they can thwart potential attacks dynamically. Advanced strategies include red teaming for AI agents and applications. This involves using AI to simulate attacks on AI systems to identify vulnerabilities proactively.

Thank you to CyberSymposiums for setting up the discussion.