Beyond alignment: Why robotic foundation models need context-aware safety

29 April 2026

Science Robotics

Alexander Robey, Zachary Ravichandran, Eliot Krzysztof Jones, Jared Perlo, Fazl Barez, Vijay Kumar, J. Zico Kolter, Hamed Hassani, and George J. Pappas, Beyond alignment: Why robotic foundation models need context-aware safety. Sci. Robot. 11, eaef2191 (2026). DOI:10.1126/scirobotics.aef2191

View Journal Article / Working Paper

Because AI-enabled robots can be tricked into taking unsafe actions, they require layered, context-aware safety guardrails.

Key Authors

Fazl Barez

Postdoctoral Research Fellow 

Other Recent Journal Article / Working Papers
Dynamics Of Sovereign Debt Credit Risk And Sustainability Analysis WP Nov 25 1

Dynamics of sovereign debt: credit risk and sustainability analysis

Publications - 31st March 2026
Wiley online library logo

Astrocyte Enrichment of 3D Cortical Constructs Enhances Brain Repair

Publications - 26th February 2026
nature

Understanding human behaviour for pandemic preparedness with epigames

Publications - 24th February 2026
Wiley online library logo

Advanced Characterization and Rejuvenation of End-Of-Life Lithium-Ion Anodes: Toward the Development of a Green Upcycling Route

Publications - 20th January 2026
nature

Stranded assets in European agriculture during food system transformations

Publications - 19th January 2026
Offshore turn to Asia

An Offshore Turn to Asia? Africa's Deepening Links with Dubai, Singapore and Hong Kong