Navigating the Frontier Between AI and Human Factors: A Three-Decade Journey in Text Input
Shumin Zhai
Principal Scientist and Engineering Director, Google
Abstract: The smartphone keyboard is one of the most used interfaces in the world, and arguably a forerunner of human-centered AI in the mass market. It is also a clear case study of a central challenge facing our field: balancing the rapid evolution of computing and AI machinery with a deep understanding of human performance, behavior and habit.
This talk uses three decades of research, design and engineering in mobile text input to examine that dynamic. It traces a line from foundational models of human performance and "laws of action" such as tapping, steering and gesturing, through the invention of gesture typing and keyboard layout optimization, to the large-scale probabilistic error correction and next word prediction serving over a billion people daily in Gboard, and most recently to AI-augmented voice writing.
As the AI frontier accelerates, accurately understanding the human side becomes more consequential than ever—both for text input and for the broader design of human-AI interaction.
Shumin Zhai is a principal scientist and engineering director at Google, leading research and engineering for Gboard, an AI-powered mobile keyboard used by over 1.3 billion daily active users across 900+ languages and varieties.
He received his BS and MS from Xidian University and his PhD in human factors engineering from the University of Toronto (1995). Following his role as a research staff member at the IBM Almaden Research Center (1996–2010), he joined Google in 2011. His research merges AI with human factors modeling, including the Law of Steering and stroke gesture production. He originated the SHARK/ShapeWriter project at IBM, pioneering touchscreen gesture typing.
Dr. Zhai's work has earned the ACM UIST Lasting Impact Award, the ACM MobileHCI Impact Award and an IEEE Computer Society Best Paper Award. He holds over 200 US patents, maintains an h-index of 88 and served as the fourth Editor-in-Chief of ACM Transactions on Computer-Human Interaction. He is an ACM Fellow (2010), a member of the CHI Academy (2010), an ACM inaugural Distinguished Scientist (2006) and an inductee into the University of Toronto’s Engineering Alumni Hall of Distinction (2020).