Skip to yearly menu bar Skip to main content

Invited Talk
Workshop: 1st Workshop on Multimodal Content Moderation

Red teaming Generative AI Systems


As generative AI systems continue to evolve, it is crucial to rigorously evaluate their robustness, safety, and potential for misuse. In this talk, we will explore the application of red teaming methodologies to assess the vulnerabilities and limitations of these cutting-edge technologies. By simulating adversarial attacks and examining system responses, we aim to uncover latent risks and propose effective countermeasures to ensure the responsible deployment of generative AI systems in new domains and modalities.

Chat is not available.