This video, presented by Tom Callahan, demonstrates how actuaries can use AI tools like Google Gemini to generate synthetic data when real-world datasets are limited. The example focuses on building 1,000 rows of synthetic claims data for a new PPO product in Chicago. While Gemini quickly produced code and incorporated relevant census data, errors and questionable assumptions showed the importance of having an “Actuary in the Loop.” The process highlights both the potential and current limitations of AI in actuarial work.
Contributors: Tom Callahan, FSA, MAAA, MS; Jon Forster, ASA, MAAA; Mason Roberts, FSA, MBA