Researcher Spotlight: Ananya Ganesh

Can artificial intelligence support multilingual learners in the classroom? Or can it unintentionally overlook the linguistic strengths they bring? As AI-powered tools become increasingly common in K–12 education, these questions become more urgent than ever. They are also the questions driving the work of Dr. Ananya Ganesh, a Postdoctoral Research Fellow at the University of Wisconsin-Madison whose research examines how language technologies can better support multilingual learners and collaborative learning in classrooms.  

This October, Dr. Ganesh will share more about her work as part of the MLRC Speaker Series. Ahead of her presentation, we asked her to reflect on what inspired her research, how her thinking has evolved, and the impact she hopes her work will have for multilingual learners, educators and researchers. 

What experiences or questions first drew you to this area of research?  

As a research engineer at ETS, I worked on assessment, specifically with automated short-answer scoring of essays. During that time, I realized how sensitive language technology is to linguistic diversity and how it could fail in unexpected or undiagnosed ways: for example, when scoring content displaying diverse accents and dialects, such as African American Vernacular English (AAVE). During my Ph.D., I developed dialog¹ to support students’ collaborative learning in K–12 classrooms and once again noticed how multilingual learners were not well-represented in either our conceptual frameworks or the data powering our models. This motivated me to both question and study the effectiveness of classroom AI tools for multilingual learners and to find ways to mitigate inadvertent failures—both of which I am grateful to explore in my postdoctoral research at UW-Madison. 

 If someone outside your field asked what you study and why it matters, what would you tell them? 

I study how technology in our classrooms can help teachers and learners communicate more effectively. Importantly, I focus on making sure that these tools help all learners equitably in modern classrooms where students come from all language backgrounds. This matters because no student should feel that their contributions are unheard or not valued simply because their mode of expression is not standard. 

What impact do you hope your work will have for multilingual learners, educators, or researchers? 

I hope my work will result in more robust systems and technology that recognize the unique contributions made by multilingual learners and surface them for teachers. For both researchers and educators, I hope that my work will help them glean actionable insights about multilingual learners’ discourse and collaboration, including during small group work. I also hope it helps them understand and improve their own discursive and pedagogical strategies to better support multilingual learners. 

Has your thinking about multilingual education changed over time? If so, how? 

My thinking about multilingual learners and multilingual education has definitely changed over time! Growing up in India, I approached multilingual learning (including my own learning) from a deficit perspective. I assumed that multilingual learners faced a barrier that must be overcome when practicing their native language. My mindset slowly changed when conducting K–12 classroom observations during my Ph.D., when I noticed how teachers in inner-city Denver schools supported multilingual learners in their classrooms by supporting differences rather than requiring conformity—for example, by providing teaching materials in multiple languages. Through working with Dr. Shamya Karumbaiah and her collaborators in the Responsible AI for Learning (TRAIL) Lab, I fully adopted an asset-oriented mindset, recognizing that learners’ fluid use of their entire linguistic repertoire is a strength that both teachers and technology must support.

1. Dialog agents are software programs designed to interact with users through conversation in a specific context or for a specific purpose. They can be designed to recognize what is happening in a conversation, for example, when a student is silent, and respond with an appropriate prompt or other intervention.

Want to hear more? 

Join us October 1, 2026, at 1 PM Central Time as Dr. Ganesh shares how her research is helping shape more inclusive educational technologies and what it means for educators and researchers. Register here.