Vigilia.
← Les avertissements
Risque d’extinction et dommages actuelsLaboratoire de pointeHead of AI Safety and Alignment, Google DeepMind4 entrées au registre

Suivre

Anca Dragan

Professeure à UC Berkeley qui dirige la sécurité et l’alignement de l’IA chez Google DeepMind depuis 2024, responsable des garde-fous de Gemini, elle présente ce travail comme couvrant à la fois les dommages actuels et les risques extrêmes liés à la progression des capacités.

Fonctions

  1. 2015 – aujourd’huiAssistant Professor, then Associate Professor, Electrical Engineering and Computer Sciences, UC BerkeleyFounded the InterACT Lab; co-PI of the Center for Human-Compatible AI; currently on leave to Google DeepMind. Start year comes from a search snippet, not a fetched page.Source →
  2. févr. 2024 – aujourd’huiHead of AI Safety and Alignment, Google DeepMindLeads the organisation formed in February 2024 to safeguard Gemini models and align future ones; also co-leads Gemini post-training. By July 2026 she signs as "VP, AI Safety & Alignment, Google".Source →

Au registre

  1. juil. 2026 · Lettre ouverte · Pacing the Frontier

    There is a real risk that capability development rapidly accelerates beyond our ability to understand or control the resulting systems.

    Head of AI Safety and Alignment, Google DeepMind à l’époque

    Signed as "VP, AI Safety & Alignment, Google"; the letter asks the US government to support international tools to deliberately pace automated AI development.

    Pacing the Frontier →

  2. 2 avr. 2025 · Publication · Google DeepMind blog

    it is essential with any technology this powerful, that even a small possibility of harm must be taken seriously and prevented.

    Head of AI Safety and Alignment, Google DeepMind à l’époque

    Lead author with Rohin Shah, Four Flynn and Shane Legg; accompanies the 30-author paper "An Approach to Technical AGI Safety and Security".

    Taking a responsible path to AGI →

  3. 28 mars 2024 · Déclaration · UC Berkeley EECS

    Google DeepMind is fairly unique in that capability and safety go hand-in-hand and have to advance together.

    Head of AI Safety and Alignment, Google DeepMind à l’époque

    Departmental announcement of her move; she adds that industry and academia "absolutely need" to work together to get safety done.

    Anca Dragan named Head of AI Safety and Alignment at Google DeepMind →

  4. 30 mai 2023 · Déclaration · Center for AI Safety

    Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.

    Assistant Professor, then Associate Professor, Electrical Engineering and Computer Sciences, UC Berkeley à l’époque

    Listed as "Associate Professor of Computer Science, UC Berkeley".

    Statement on AI Risk →