Thoughtful AI: Designing AI as a Companion for Learning and Inclusion

Описание: Today, AI systems are becoming faster and more convenient, but they often replace reflection with quick answers. In this talk, I argue for a different approach: Thoughtful AI, where AI systems are designed to help people think, learn, and reflect rather than simply respond. Thoughtful AI centers care, slows interaction, and keeps humans in the loop, especially in settings where formal support systems are limited.In this talk, I discuss my PhD research on a collaborative chatbot for rural teachers in Côte d’Ivoire. In a year-long deployment, the system supported peer learning among over 400 teachers and indirectly impacted more than 10,000 students, showing that community-centered designs can outperform traditional chatbots. I then turn to my postdoctoral work, where I explore how generative AI can help Afghan women safely learn programming outside formal classrooms, and how large language models can improve survey clarity and fairness across cultural contexts.I conclude by outlining my future research agenda, which extends Thoughtful AI to new domains of public impact, including civic participation, health, and sustainability. Across these projects, I ask a central question: How can AI help people think more—not less—where human expertise and institutional support are scarce?Bio: Vikram Kamath Cannanure is a postdoctoral fellow at Saarland University, where his research advances AI systems that center care, slow interaction, and keep humans in the loop—for communities often excluded from meaningful learning and decision-making. His work bridges Human–Computer Interaction (HCI) and Artificial Intelligence, with a focus on participatory design and low-resource contexts.He earned his PhD in Human–Computer Interaction from Carnegie Mellon University, where he designed and deployed collaborative chatbots to support teacher professional development in rural Côte d’Ivoire, reaching over 400 teachers and impacting more than 10,000 students. In his postdoctoral work, he explores how generative AI can support Afghan women in learning programming outside formal classrooms and develops LLM-powered tools to improve survey clarity and fairness across cultures. His research spans India, Rwanda, Ivory Coast, Afghanistan, Cameroon, South Africa, Macau, Germany, and the United States, and has been published at venues including CHI, ICTD, COMPASS, and PDC.