Practical AI Governance for Sentients

I wanted to build a simple app to help my kids’ high school run a fundraiser. I didn’t have years to turn myself into a full-time software developer first, so I partnered with an AI co-pilot and started vibe coding my way from an idea to a working product in nine months. That “simple app” quickly turned into GradVids, a production platform handling video uploads, SMS and email delivery, admin roles, and real users on the Internet while countering bot attacks. As the app moved from a proof of concept into AWS and then into production, the pace was unlike anything I had experienced before, and the risks were very real. While features were shipping quickly, bots were probing for weaknesses, assumptions were being tested, and AI-generated code was confidently solving problems without ever considering an adversary. This talk is a candid, real-world case study of what happens when vibe coding meets production reality. Using GradVids as the setting, we will explore where AI dramatically accelerated development, where it quietly introduced security and compliance risk, and why human judgment is still essential for threat modeling, trust boundaries, and guardrails. Attendees will leave with practical lessons for using AI as a force multiplier, without burning the house down.