Dr. Fei-Fei Li highlights spatial intelligence as the next frontier for AI, moving beyond language models focused on words. Spatial intelligence allows AI to perceive, reason about, and interact with the 3D physical and virtual world, the core of human cognition. This new capability helps AI not just describe but navigate and create within real environments.
Spatial intelligence builds on how humans learn about the world through physical interaction and spatial reasoning, essential from childhood learning to expert tasks. At World Labs, Dr. Li works on “world models,” advanced generative AI systems that create spatially consistent 3D worlds obeying physical laws and dynamics, unlike current models limited to text and images.
These world models integrate multimodal data like images, videos, and actions to predict real-world changes over time, giving AI the agility and imagination to navigate, create, and reason scientifically. This approach aims to shift AI from “wordsmiths in the dark” to intelligent agents deeply aware of spatial context and capable of meaningful interaction.
Spatial intelligence will transform industries from creativity and storytelling to robotics and health, making machines truly spatially aware and interactive. It represents a profound leap in AI’s ability to integrate imagination, perception, and action, opening new horizons for machine-human collaboration.
THE FUTURE OF AI IS WORLDS, NOT JUST WORDS.