We have seen tens of thousands of posts, blogs, articles, and more about the threat of ‘Artificial Intelligence’ in the news – and the reality is that many companies, from Microsoft and Google down to small App Dev houses and Academic Researchers, are exploring how to integrate this technology into applications… so for better or worse, it’s coming!
This talk will explore how to hack these applications, not just ‘prompt hacking’ the AI to make it misbehave, but actively using a LLM-based chatbot to do you hacking for you! We explore the various ways you can compromise the app, start to develop a threat model for including these bots in your apps, as well as give an outline of protections and proof-of-concept code to begin to defend these apps from attack.
leHACK website: https://www.lehack.org
leHACK's twitter: https://twitter.com/_lehack_
leHACK instagram : https://www.instagram.com/_lehack_