Fundamental AI Research
We develop the AI foundations needed to understand complex human and biological systems.
This vertical focuses on how models learn safely, robustly, and collaboratively over time.
Our current efforts are organized into two themes:
- AI Safety & Trustworthiness – bias, uncertainty, robustness, demographic leakage, and secure behavior.
- Human–AI Ecosystems – multi-agent intelligence, lifelong learning, and experience sharing (SheLL).
You can explore each theme in more detail below.