Research Seminar by Chen Jingjing | Advancing Responsible Generative AI: Progress and Challenges

Advancing Responsible Generative AI: Progress and Challenges

In recent years, generative artificial intelligence has made significant progress in visual content generation. However, alongside the generation of high-quality visual content, there arise new challenges in ensuring the security and regulatory compliance of visual models. This talk delves into the ongoing progress and challenges in advancing responsible generative AI, specifically focusing on visual content creation. Additionally, we will share our latest work in building responsible generative AI, including frameworks for generalizable AI-generated visual content detection in open environment, and approaches for achieving reliable erasure of prohibited concepts without compromising the model's generation capabilities. Through this discussion, we aim to contribute to the ongoing development of ethically sound and socially responsible generative AI systems.