Attacks on and defences for machine learning models in production.
An agent's blast radius is not what it was asked to do. It is everything its credentials permit, available to anyone who can influence its input. Four changes that shrink it.
·6 min read
There are plenty of AI DLP solutions out there (which is what we expected), but there is currently nothing that controls the commands an AI agent can or cannot execute, and that is fundamental. Think about this: An employee says to the AI...
·3 min read
Your internal assistant reads the wiki and can raise tickets. That means anyone who can edit the wiki can make it act. Here is exactly how that plays out, and the four changes that stop it.
·6 min read ·1 like