Hen Davidov is a Rhodes scholar leveraging artificial intelligence in a pursuit to transform healthcare and data-informed decision making. Graduating with distinction in Computer Science from the Technion, Hen was also honored with the Israeli Planning and Budgeting Committee Awards for Data Science in 2023 and 2024.
Hen’s research focuses on developing reliable and safe AI models to enhance decision-making in high risk settings. During Hen’s Master’s work, supervised by Prof. Yaniv Romano, Prof. Ron Kimmel, and Dr. Gil Shamai, he created statistical wrappers that provide guaranteed, interpretable outputs from prognostic deep learning models, giving healthcare professionals greater confidence in AI-assisted prognoses. In another work, he proposed a novel measure of LLM safety, time-to-unsafe-sampling – the amount of times a given prompt can be used as input for an LLM, before triggering an unsafe response. Surprisingly, survival analysis tools can be used for trustworthy prediction of this safety quantity.
