Speaker

Naimul Hoque

Abstract

AI-infused supertools are applications that amplify, augment, empower, and enhance human performance, by combining user experiences with AI support services [1]. In the past decades, we have made progress towards this vision of AI supertools. Generative AIs such as GPT-4 and DALL-E3 can generate human-like text, converse with humans, and produce futuristic images that amplify human creativity and imagination. However, at the same time, we see concerns about generative AIs threatening to overtake human agency and ownership. For example, the Writers Guild of America (WGA) and the Screen Actors Guild - American Federation of Television and Radio Artists (SAG-AFTRA), were recently on strike, demanding clauses on their contract that they will not be replaced by AI.

In this talk, I will discuss how interactive visualization can be a key enabling technology for retaining human agency and ownership in AI supertools. Interactive visualization can work as a communication medium between humans and AI that people can use to steer AI towards their needs while having clear control over the final artifact (e.g., novel, code, etc.). To validate this premise, I will present three AI-assisted writing tools to show how interactive visualization and AI can amplify creative writing while retaining writers’ agency and ownership. AI-assisted writing is a controversial area with high tensions between AI and humans, which makes the domain a suitable probe for validating my research agenda. I will conclude by summarizing lessons learned from the design process of the tools and proposing future works to cement Designing AI-Infused Supertools with Interactive Visualization as an emerging research area for the HCI and visualization community.

[1] Ben Shneiderman. 2022. Human-Centered AI. Oxford University Press, Oxford, United Kingdom.

Bio

Naimul Hoque is a PhD candidate at the College of Information Studies (iSchool), University of Maryland, College Park. He works with Professor Niklas Elmqvist in the Human-Computer Interaction Lab (HCIL), one of the oldest and most well-known HCI research labs in the U.S. His research focuses on designing, building, and validating AI-infused applications using interactive visualization. This typically results in novel data visualization and interactive systems, HCI theories and methods, and AI applications. Naimul has published works at top-tier HCI and visualization venues such as ACM CHI, IEEE VIS, ACM CSCW, and ACM DIS during his PhD. He obtained a Master's in Computer Science from Stony Brook University in 2020 and a B.Sc in Computer Science from the University of Dhaka, Bangladesh in 2015.