In this talk, we’ll dive into the fascinating and critical topic of hacking large language model (LLM) chatbots and AI applications. As these systems become integral to industries like customer support, education, and software automation, understanding their vulnerabilities is essential. We'll explore real-world examples of exploits, including prompt injection, adversarial manipulation, and API abuse, demonstrating how attackers can manipulate AI behavior or extract sensitive information. Additionally, we'll discuss strategies to harden these systems against threats, from input validation to adversarial training. This session will leave you with a deeper understanding of the risks, as well as actionable insights to protect and secure AI-driven applications.