Many thanks to the UKRI AI Hub in Generative Models for featuring our work and for their support of this research.

Gen AI - the AI Hub in Generative Models@genaihub.bsky.social · 5mo ago

A method for testing AI safety by using human-like concepts to trick generative models into making mistakes, has been developed by Hub researchers. The paper, entitled Concept-based Adversarial Attack: A Probabilistic Perspective, is summarised on our website: www.genai.ac.uk/news/hub-res...

A method for testing AI safety by using human-like concepts to trick generative models into making mistakes, has been developed by Hub researchers. The paper, entitled Concept-based Adversarial Attack: A Probabilistic Perspective, is summarised on our website: www.genai.ac.uk/news/hub-res...