[general_dat] Seminario AISAR – Safeguarded AI: a scalable workflow for safety-by-construction

[Agustín Martinez Suñé]

Desde el Programa de Becas AISAR en AI Safety tenemos el placer de
invitarlos a la próxima charla de nuestro seminario online, con la
participación de investigadores del área.

📌 Fecha y hora: Miércoles 5 de noviembre, 11:00 hs (ARG).
🎤 Oradora: Nora Ammann – Technical Specialist @ UK's Advanced Research and
Invention Agency
📖 Título: Safeguarded AI: a scalable workflow for safety-by-construction
🔗 Charla online: Para asistir a la charla, registrate acá:
https://luma.com/mkvikgf4

Abstract: This talk introduces Safeguarded AI, an ambitious, $70M R&D
programme developing a general-purpose workflow for safety-by-construction:
a method for leveraging frontier AI capabilities to build domain-specific
AI systems with quantitative safety guarantees. We explore how this could
unlock the benefits of advanced AI for deployment in safety-critical
settings like power grid balancing, supply chain management, and clinical
trial design, and argue that this form of 'strategic narrowness' may be key
to unlocking safe and transformative AI. Finally, we will situate
Safeguarded AI within a broader family of approaches—known as Guaranteed
Safe AI—that combine structured world models, formal safety specifications,
and compositional verification to automate the construction of AI
applications with quantitative guarantees of safety.

Más información:
https://www.aria.org.uk/opportunity-spaces/mathematics-for-safe-ai/safeguarded-ai/
Equipo AISAR
http://scholarship.aisafety.ar/
<http://scholarship.aisafety.ar/?utm_source=chatgpt.com>