The Turing Lectures: Frontier AI Under Pressure — Building Resilience (Highlights)

Frontier AI is evolving faster than our institutions can keep up. How do we build systems that don't just react to change, but are truly resilient? In this highlight from the 2026 Turing Lectures, Dr. Mikel Rodriguez—leader of the Department of Defence’s AI Red Team and co-creator of the ATLAS framework - reveals the pressure points where rapid AI advancement meets our most fragile institutions. This isn't about perfect control. It’s about navigating the trade-offs between innovation and vulnerability. Dr. Rodriguez draws on his frontline experience at Google DeepMind and MITRE Labs to explain why the future of resilience relies on humility, diversity, and decisive action amid uncertainty. Key Takeaways - Why frontier AI (like GPT-5/Gemini) is outpacing institutional policy. - The breakdown between technical security and human decision-making. - How to build systems that adapt to the next generation of adversarial threats. - Moving beyond the "control" mindset in AI safety. 🔗 Watch the full lecture here: https://www.youtube.com/live/s-O91U8gfNQ?si=zl_iaZfI2CF44zL- 📢 Explore the 2026 Turing Lecture Series: https://www.turing.ac.uk/events/the-turing-lectures 00:00 Welcome and Theme 00:52 Can Machines think? 03:17 Meet Dr Mikel 04:05 Early Failures and Winters 05:32 Dormant, Then sudden - Deep Learning Breakthrough 08:51 Institutions Lag Behind 10:37 Pressure Point 1 Cyber Dual Use 12:33 Pressure Point 2 Proliferation 14:59 Pressure Point 3 Counter AI 18:20 Pressure Point 4 Monoculture 19:51 Pressure Point 5 Asymmetry 21:14 Pressure Point 6 Policy Velocity 21:54 Unifying Pattern Across Gaps 23:08 Defining Resilience Margin 24:58 Geopolitics and Regulation