Ann Arbor, Michigan, USA: Angana Borah, an Indian-origin doctoral candidate from Assam, has been awarded the prestigious Rackham Predoctoral Fellowship on April 8, 2026, in recognition of her significant contributions to artificial intelligence research at the University of Michigan.
Currently pursuing her Ph.D. in computer science and engineering, Borah’s research focuses on enhancing large language models to promote cultural inclusivity, factual accuracy, and broader social well-being. Her work aims to address systemic biases in AI systems and improve how these technologies interact across diverse cultural contexts.
The Rackham Predoctoral Fellowship is awarded annually to a select group of doctoral candidates who demonstrate exceptional progress toward completing their dissertations. The selection process evaluates academic achievements, including publications, conference presentations, and contributions to teaching, mentorship, and service.
Borah’s research tackles a critical challenge in modern AI development. Many existing language models are trained predominantly on data representing Western, educated, industrialized, rich, and democratic populations. This imbalance often leads to systems that overlook regional nuances, reinforce stereotypes, or exhibit inconsistent behavior in group-based interactions.
Her work investigates these limitations through multi-agent AI systems and socially grounded evaluation methods. In one of her studies, Borah analyzed interactions between multiple AI agents, revealing that biases can re-emerge and even intensify during collaborative interactions, despite individual models appearing unbiased.
In another project, she explored cross-cultural image captioning by developing a framework called MosAIC. This approach leverages agents with diverse cultural perspectives to generate more context-aware and inclusive descriptions. The research also introduced new datasets and evaluation metrics to better assess cultural representation in AI outputs.
Her findings have been presented at leading international conferences, including ACL, EMNLP, NAACL, EACL, and AAAI, highlighting the global relevance of her work.
Praising her contributions, her advisor Rada Mihalcea noted that Borah has demonstrated exceptional capability in identifying globally significant research challenges while delivering impactful and practical solutions at an early stage in her academic career.
Beyond research, Borah has actively contributed to academic and global initiatives. She has co-organized workshops such as “NLP for Positive Impact” at ACL 2025 and participated in teaching and outreach programs in countries including Kenya and Romania.
The fellowship will support her continued efforts to develop AI systems that are more equitable, culturally aware, and socially responsible.
