Mandatory human-in-the-loop is a cybersecurity cop-out. People are giving agents more and more autonomy. We need solutions that accept that world because there is no stopping it.
It's like telling people in the 90s to not use the internet to avoid getting hacked. Good luck.
I think EVERYONE doing AI security should watch this video. It's a great model for understanding the risks various levels of agency we give to our AIs
It seems like such an obviously good idea to have a human in the loop until you realize it's just going to be: "Press F1 to continue" over and over.Join the conversation
