Throughout my career, I’ve found myself fascinated by statistics. Back in high school and college, the subject felt arcane and unintuitive, with rote memorization of how to use probability distribution lookup tables and expressing values as Z-scores. Launching my career in supply chain management and operations research deepened my appreciation for the science of uncertainty and how data can help quantify it. Even more fascinating was how human behavior created and reacted to information, and how easy it was to misunderstand data and findings. Teaching Artificial Intelligence System Safety at the University of Southern California gave me an opportunity to dive deep into these problems. Sharing my findings with students from all over the world and from various government agencies allowed me to demonstrate a universal truth: new problems are usually just old problems in disguise.
My interest in statistics accelerated with the renaissance of artificial intelligence alongside the data science boom in the early 2010s. From basic machine learning to large language models, I was struck by how deeply statistics drove these innovations and how the AI boom amplified both the strengths and flaws of statistical analysis. I saw this firsthand while working for airlines, the tech sector, my own two startups, and the university where I regularly teach.