Why AI needs Human-Centered Design

The Informatics Seminar Series is proud to present James Guszcza, Deloitte.

Recorded on Friday, April 19, 2019.

Talk: “Why AI needs Human-Centered Design”

Modern AI is less about creating human-like general intelligence than it is about creating tools that do cognitive spade work and more generally, enhancing or extending human intelligence. AI tools based on statistical learning, big data, and pattern recognition can perform a growing number of tasks that are difficult or impossible for humans to perform. However, they perform poorly at many aspects of cognition that come naturally to humans: formulating hypotheses, understanding cause and effect relationships, using commonsense reasoning, picking up on social cues and nonverbal forms of communication, and expressing empathy.

The complementary nature of human and algorithmic intelligence points to the need for an interdisciplinary approach that draws on such fields as computer science, human psychology, behavioral economics, design thinking, and ethics: designing collaboration systems that enable forms of human-computer collective intelligence. Jim will sketch a framework to organize different types of AI projects, and offer real-life examples in which human-centered design has been crucial to the economic success of an AI project. Concepts covered will relate to both System 2 cognition (“thinking slow”) and System 1 cognition (“thinking fast”). Regarding the former, JCR Licklider’s notion of human-computer symbiosis is relevant: algorithms are good at what humans are poor at and vice versa. Regarding the latter, behavioral economics teaches us that prompting smarter choices and decisions often involves more than providing information or setting up incentives. Often the way information is presented or choices are arranged has surprisingly large effects on end-user behavior. Thus AI systems will often benefit from insightful uses of choice architecture. James shares a number of AI examples to illustrate these principles.

Jim is the US chief data scientist of Deloitte Consulting and a member of Deloitte’s Human Capital practice. Jim has 20 years of experience-building and designing analytical solutions in a variety of public and private sector domains. In recent years, he has spearheaded Deloitte’s use of behavioral nudge tactics to more effectively act on algorithmic indications and prompt behavior change. Jim is a former professor at the University of Wisconsin-Madison business school, and holds a PhD in the Philosophy of Science from The University of Chicago. In addition, Jim is a Fellow of the Casualty Actuarial Society and recently served on its board of directors.