Distorting or enhancing realities using Generative AI

Go back to programme

Distorting or enhancing realities using Generative AI

In this presentation, Hao Li will talk about the impact of Generative AI on digital humans research and how they have evolved from computer generated avatars to photorealistic deepfakes. He will highlight existing and new emerging capabilities for generative AI, and present GAN-and Diffusion-based image and video synthesis methods. He will show examples of real world threats as well as present a number of positive applications, especially those done at my startup company, Pinscreen, where they applied these AI technologies for Hollywood. In particular, he will show cutting edge examples in visual effects where we used Generative AI for face replacement, De-Aging, and AI lipsyncing for vsual dubbing. Finally, he will present our latest research at the MBZUAI Metaverse Center in virtual reality, where Generative AI is used to create accessible and photorealistic avatars for immersive telepresence.

Share this session

Are you sure you want to remove this speaker?