AI security risk is national security risk
Автор: StackAware
Загружено: 2024-02-07
Просмотров: 98
Описание:
I was honored to brief the bipartisan staff of the United States Senate Select Committee on Intelligence.
Here, I discuss unintended AI training and why it's not new, but generative AI supercharges it.
We've been training predictive AI models for a while now to get better at:
Water pressure monitoring
Spam detection
Credit scoring
Some of these things can be sensitive, but generative AI tools like Large Language Models (LLM) bring things to a whole new level.
That's because people provide them MUCH more information.
And they can be trained on it.
Which is something both government and corporate security teams should stay alert for.
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке:
