Видео с ютуба Ai Security Institute

Same Model, Same Benchmark, 42% vs 95% — What Went Wrong? | Dr. Cozmin Ududec, AI Security Institute

Sandbox breakout evals with Inspect - UK AISI @ FC London '25

New Paper Explores the Flaws in AI Safety Research

AI safety: OpenAI, Meta, Google, and more face the Business, Innovation, Science and Trade Committee

Отчет об инциденте Института безопасности ИИ № INC-2026-07-28-01

AI systems are growing faster than safeguards, warns AI expert Helen Toner | 7.30

When AI Lies: The UK AI Security Institute Deception Incident

Anthropic's Mythos, the UK AI Security Institute, and the Dawn of Autonomous Cyber Infiltration

I Run the World’s Oldest AI Safety Lab... There’s a “High Double-Digit” Chance AI Kills Us

Джеффри Хинтон о рисках для безопасности ИИ и будущем ИИ | IASEAI '26

Обзор курса — Безопасность ИИ

A.I. Security Institute hacking incident reviewed.

UK AI Security Institute Opens More Model Evaluation Results

AISI's AI Agents Attacked Real People - The Third Disclosure in Three Weeks

UK AI Security Institute: 19 unsanctioned AI agent actions

AISI AI Agent Test goes rouge

Case study: Isambard-AI: AI Security Institute

Mythos and the New Cyber Frontier Masterclass

Агенты OpenAI и Anthropic демонстрируют новое «несанкционированное» поведение ИИ

15-минутная еженедельная привычка в сфере ИИ — Институт безопасности ИИ