Nightshade: Prompt-Specific Poisoning Attacks on Text-to-Image Generative Models - Session 335

Nightshade: Prompt-Specific Poisoning Attacks on Text-to-Image Generative Models - Session 335

Explore a cutting-edge research presentation on prompt-specific poisoning attacks targeting text-to-image generative models. Delve into the innovative "Nightshade" technique, which introduces a novel approach to manipulating AI-generated images. Gain insights into the potential vulnerabilities of popular image generation models and understand the implications for AI security and ethics. Learn about the methodology, experimental results, and potential countermeasures discussed by researcher Shawn Shan in this thought-provoking IEEE conference talk.