Agentic AI Security in One Sentence (and Why Boards Care)
Автор: David Campbell
Загружено: 2026-02-09
Просмотров: 520
Описание:
If you remember one sentence from this video, make it this:
An agent is software that can be steered by text into taking actions.
That’s why agentic AI security isn’t a vibes conversation about “model safety.” It’s system security.
In this short episode, I explain agentic AI security in board-level terms:
'What makes an AI system agentic'
'Why the security boundary shifted from the model to the action supply chain'
'How a single malicious snippet can trigger real-world impact'
'What “good” controls actually look like in practice'
This framing is designed for executives, board members, and security leaders who need clarity without jargon.
Next episode: Prompt injection explained as a trust-boundary problem, not a “prompt problem.”
References
NIST AI 600-1: Generative AI Profile
https://nvlpubs.nist.gov/nistpubs/ai/...
OWASP AI Agent Security Cheat Sheet
https://cheatsheetseries.owasp.org/ch...
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке:
