Jade Leung

Directrice technique de l’AI Security Institute britannique et, depuis août 2025, conseillère IA du Premier ministre, elle dirige les évaluations publiques des modèles de pointe pour les risques cyber, biologiques et de perte de contrôle, et soutient que ce sont des organismes publics, et non les développeurs, qui doivent les juger.

Fonctions

2018 – aujourd’huiCo-founder and Head of Research and Partnerships, Centre for the Governance of AI, University of OxfordStart year not independently verified.Source →
2021 – oct. 2023Governance Lead, OpenAISource →
oct. 2023 – aujourd’huiChief Technology Officer, UK AI Safety Institute (renamed AI Security Institute in February 2025)Founding CTO; designs and oversees pre-deployment evaluations of frontier models.Source →
août 2025 – aujourd’huiAI Adviser to the Prime Minister, 10 Downing StreetSucceeded Matt Clifford; splits her time between Number 10 and the AI Security Institute.Source →

Au registre

nov. 2024 · Publication · AI Safety Institute blog
These evaluations give governments an insight into the risks developing at the frontier of AI, and an empirical basis to decide if, when, and how to intervene.
Chief Technology Officer, UK AI Safety Institute (renamed AI Security Institute in February 2025) à l’époque
Co-written with AISI director Oliver Illott in a November 2024 blog post; quoted verbatim by TIME on 16 January 2025.
Inside the U.K.'s Bold Experiment in AI Safety →
5 sept. 2024 · Entretien · TIME
You really want a public interest body that is genuinely representing people to be making those decisions.
Chief Technology Officer, UK AI Safety Institute (renamed AI Security Institute in February 2025) à l’époque
On who should decide whether a frontier model is safe to release.
Jade Leung: The 100 Most Influential People in AI 2024 →