Securing AI Part 3: AI Agents - Use Cases and Security

A10 security experts, Jamison Utter, Diptanshu Purwar, and Madhav Aggarwal explore the topic of securing AI agents, which they define as systems that perceive, decide, and act.
They discuss:
Defining AI Agents: Explaining that agents are not just chatbots, but are the "hands of AI" that can execute actions, call APIs, and automate complex workflows.
The Challenge of Security: Discussing how security for AI agents goes beyond traditional model security and includes protecting against prompt injection, malicious instructions, and preventing unsafe actions or data leakage.
The Importance of Context and Data: Highlighting the critical role of verifiable signals and well-defined, concise task specifications. They note that a lack of high-quality data and feedback can lead to "task drift," where the agent's actions deviate from its intended purpose.
Governance and the Role of Networking: Emphasizing the need for strong governance frameworks, including external controls like AI firewalls and DLP (Data Loss Prevention), to ensure accountability and policy alignment. They conclude by discussing the emerging opportunity for networking companies to provide these critical security and governance solutions for the new age of AI.
Learn how to secure AI and LLMs in your organization: https://bit.ly/4kOHmYd
#aiagents #aisecurity #guardrails #cybersecurity2025 #aifirewalls #a10networks