Assistant professor Fan Yao uses statistical learning to study how humans and artificial intelligence-driven applications interact. When Fan Yao was a Ph.D. student at the University of Virginia, he had the opportunity to intern with Meta. Yao’s thesis research revolved around the responsible design of online platforms, including recommendation algorithms like the ones that provide endless streams of social media content based on your scrolling habits. During his internship, Yao co-designed a mechanism that, in a long-term experiment on Instagram Reels, measurably improved user engagement and the diversity of content users consumed. Recommendation systems often prioritize short-term relevance and engagement, which can incentivize creators to produce content for users who are already well-served on the platform. The mechanism designed by Yao and his collaborators directed more attention toward underserved users, improving engagement while encouraging a more diverse content ecosystem. “I don’t see the advancement of technology as a purely technical issue,” Yao said. “The advancement of technology should always serve people and society.”