We are now the AI Security Institute
Please enable javascript for this website.

Safety Cases: A scalable approach to Frontier AI safety

Read the Full Paper

Authors

Benjamin Hilton
Marie Buhl
Tomek Korbak
Geoffrey Irving

Benjamin Hilton, Marie Buhl, Tomek Korbak, Geoffrey Irving

Abstract

Discusses the use of safety cases in frontier AI development, arguing they can help demonstrate AI safety to decision-makers. It also highlights how safety cases can support the Frontier AI Safety Commitments and outlines open research questions on their methodology and implementation.

Notes