Evaluating AI Security: Performance vs. Safety
Автор: A10 Networks
Загружено: 2025-07-31
Просмотров: 714
Описание:
Evaluating AI Security: Performance vs. Safety
In this video, A10 Networks' security leaders Jamison Utter, Madhav Aggarwal, and Diptanshu Purwar discuss the crucial considerations for evaluating AI security within an organization.
Madhav Aggarwal emphasizes the following points:
Evaluating AI Safety: The way to assess AI at your organization is to verify that systems are secure, which involves implementing various "guardrails" designed to operate on top of AI.
Performance vs. Safety: A critical insight is that a highly performant AI system may not inherently be safe. Performance and safety are distinct aspects that do not automatically align.
A Frontier for AI Companies: The challenge of simultaneously achieving both high performance and robust safety is a significant hurdle for trending
AI companies operating in this evolving frontier. It is often difficult to attain both through the same AI model.
This segment delves into why achieving both can be a complex task through the same model.
Learn more about our threat intelligence platform and securing AI and LLMs: https://bit.ly/4kOHmYd
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке:
