Securing AI Agents: When Your LLM Has Tool Access

Modern AI agents have evolved beyond simple text processing; they are now capable of autonomous action, from browsing and creating content to executing complex commands. However, this power comes with significant risk, as subtle adversarial prompts can subvert their function, turning a trusted assistant into a vector for attack.
Join this session to delve into the security implications of autonomous agents, exploring real-world attack scenarios and effective defense strategies in this rapidly changing landscape.