Understanding Generative AI with Stefanie Mueller

MIT Associate Professor Stefanie Mueller is PI of the HCI Engineering Group (Human-Computer Interaction) at CSAIL. Her group leads groundbreaking research using generative AI for 3D applications. In this episode she explains how generative AI can be combined with mechanical simulation to create stable and personalized 3D models.
Generative AI can already create stunning images and compelling text, but ask it to design a coffee mug you can actually use and things start to fall apart. Handles snap, walls get too thin, and physics can be entirely ignored. Mueller explains why today’s AI struggles with physical functionality and how her team is tackling that problem by combining generative AI with mechanical simulation.
You’ll hear about Mech-Style, an approach that allows AI to personalize 3D models—like adding textures or stylistic patterns—while ensuring the final object remains structurally sound and printable. From customized consumer products to the future of mass personalization through 3D printing, this conversation dives into how AI could fundamentally change the way things are designed and manufactured.
 

 Find a full transcript of this podcast here.