09 October 2025 14:30 - 15:00
Panel | Ensuring model safety: Approaches to secure & robust AI in critical applications
In high-stakes domains, AI systems must be both secure and resilient to unpredictable conditions.
This panel brings together experts from industry and research to share approaches for safeguarding models against adversarial threats, ensuring robustness under stress, and aligning outputs with intended objectives.
Panelists will explore practical frameworks, continuous evaluation methods, and governance strategies that keep AI reliable in critical real-world applications.