Agenda Pushing In Email To Thwart Phishing

In this work, we draw parallels in automatically responding to emails for combating social-engineering attacks and documentgrounded response generation. We lay out the blueprint of our approach and illustrate our reasoning. E-mails are longer than dialogue utterances and often contain multiple intents. To respond to phishing emails, we need to make decisions similar to those for document-grounded responses-deciding what parts of long text to use and how to address each intent to generate a knowledgeable multi-component response that pushes scammers towards agendas. We propose Puppeteer as a promising solution to this end: a hybrid system that uses customizable probabilistic finite state transducers to orchestrate pushing agendas coupled with neural dialogue systems that generate responses to unexpected prompts. We emphasize the need for this system by highlighting each component's strengths and weaknesses and show how they complement each other.