I study how AI mediates human communication across languages and cultures, and how humans, in turn, learn to interpret and make decisions from imperfect AI outputs. My research sits at the intersection of multilingual NLP and human-centered AI, where I combine methods from NLP, HCI, and statistics to design and evaluate multilingual systems and to measure their impact on human understanding and decision-making. My long-term goal is to make AI-mediated communication more trustworthy for diverse communities.