
When a user interacts with an agent, the documents and discussion enter a model’s context and can influence subsequent actions. As Arist’s Mike Doty puts it: “Our threat model has changed from ‘what can a user send us?’ to ‘what can a user get the system to do?’”












