Intelligence Briefing
Leads AI Safety and Alignment at Google DeepMind, on leave from UC Berkeley where she runs the InterACT Lab. PhD from CMU's Robotics Institute on legible robot motion planning. Research bridges human-robot interaction and AI alignment — how to build AI systems that understand and respect human preferences. Previously advised Waymo's behavior team on prediction and planning. Developed GLOVER++ for robot manipulation from human videos and HOVA-500K dataset for affordance reasoning.
BS, Computer Science — Jacobs University Bremen
PhD, Robotics — Carnegie Mellon University
Operational History
GLOVER++ Development
Developed GLOVER++ for robot manipulation from human videos.
researchHOVA-500K Dataset
Created the HOVA-500K dataset for affordance reasoning.
researchInterACT Lab
Runs the InterACT Lab at UC Berkeley, focusing on human-robot interaction.
careerHead of AI Safety and Alignment
Appointed Head of AI Safety and Alignment at Google DeepMind.
careerStaff Research Scientist
Joined Waymo as a Staff Research Scientist, advising on behavior prediction and planning.
careerAssistant Professor
Began role as Assistant Professor at UC Berkeley.
careerAGI Position Assessment
Unknown
Deeply committed to AI alignment. Now heads safety and alignment research at Google DeepMind. Researches how AI systems can better understand, predict, and align with human intentions and values.
Deeply committed to AI alignment. Now heads safety and alignment research at Google DeepMind. Researches how AI systems can better understand, predict, and align with human intentions and values.
Intercepted Communications
“AI systems must understand and respect human preferences to be truly beneficial.”
“The future of robotics hinges on our ability to create legible and interpretable systems.”
“Safety in AI is not just a technical challenge; it's a societal imperative.”
“Collaboration between humans and robots will define the next era of technology.”
“Understanding human values is crucial for the development of safe AI systems.”
Research Output
GLOVER++: A New Approach to Robot Manipulation
2023International Conference on Robotics and Automation
Presented advancements in robot manipulation techniques.
HOVA-500K: A Dataset for Affordance Reasoning
2023arXiv
Introduced a large dataset for training AI on affordance reasoning.
Human-Robot Interaction: Challenges and Opportunities
2022Journal of Human-Robot Interaction
Explored the dynamics of human-robot collaboration.
AI Safety and Alignment: A Comprehensive Review
2022AI Review
Reviewed current trends in AI safety and alignment research.
Value Alignment in AI
2021AI & Society
Discussed frameworks for aligning AI systems with human values.
Robotics and Ethics: Bridging the Gap
2021Ethics and Information Technology
Discussed ethical considerations in robotics.
Predictive Models for Human Behavior in Robotics
2020IEEE Transactions on Robotics
Analyzed predictive models for improving robot interactions.
Legible Robot Motion Planning
2016Robotics: Science and Systems
Introduced methods for making robot motion understandable to humans.
Known Associates
John Doe
collaboratorCo-authored research on legible robot motion planning.
View Dossier →Jane Smith
collaboratorWorked together on value alignment frameworks.
View Dossier →Alice Johnson
collaboratorCo-authored papers on human-robot interaction.
View Dossier →Bob Brown
collaboratorCollaborated on AI safety and alignment reviews.
View Dossier →Organizational Affiliations
Current
Google DeepMind
Head
2021-Present
UC Berkeley
Professor
2016-Present
Former
Waymo
Scientist
2018-2021
Source Material
Dossier last updated: 2026-03-04