Secure Generative AI Agents Workshop

Goal

The workshop will investigate the safety, security, and privacy of GenAI agents from a system design perspective. We believe that this new category of important and critical system components requires a whole new approach and we intend for this workshop to explore new kinds of security and privacy for AI agents. We aim for this year’s workshop to inform, through a vigorous discussion, the research agenda of the security and privacy community around GenAI agents.

Background

Generative AI (GenAI) agents are quickly becoming the preferred mechanism for GenAI deployments, with predictions that they will become as transformational as the Internet and Smartphones. At its core, a GenAI agent relies on machine-learning (ML) models trained on large amounts of data using deep-learning techniques. AI agents make autonomous use of such models, often rely on reasoning to sequence tasks towards a goal, and call into external (non-ML) components such as databases, APIs, and tools. Agentic designs have significant security and privacy challenges (some inherited from the underlying models, see https://genai.owasp.org/llm-top-10/ ), which can be addressed only in a holistic manner that considers the AI agents together with the (non-ML) systems built around them.

Workshop Information

The workshop is part of the 46th IEEE Symposium on Security and Privacy workshop series, and will take place in San Francisco, CA, on May 15, 2024.

For more information, see:

• workshop program

• workshop venue