AI and HCI: Two Fields Divided by a Common Focus

48 AI MAGAZINE AI and human-computer interaction (HCI) are converging. “Usable AI” conference events in 2008 and 2009 preceded this special issue, and ACM will launch a widely-supported Transactions on Interactive Intelligent Systems. AI techniques are in the toolset of more and more HCI researchers, and applications of machine learning are increasingly visible in the HCI literature. Other maturing AI technologies seek input from the HCI community. The two fields have met under shared tents for some time, notably within International Journal of Man-Machine Studies (subsequently International Journal of Human-Computer Studies) and at the Intelligent User Interface conferences cosponsored by ACM’s Special Interest Groups on Computer-Human Interaction (SIGCHI) and Artificial Intelligence (SIGART). But little of this research has flowed back to the major AI and HCI conferences and journals. In this article, I describe some research that has bridged the fields, but contact has been sporadic. Logically, they could have been closer. Both explore the nexus of computing and intelligent behavior. Both claim Allen Newell and Herb Simon as founding figures. Working over the years as an HCI person in AI groups at Wang Laboratories, MIT, MCC, and Microsoft, and alongside AI faculty at Aarhus University and the University of California, Irvine, I was puzzled by the separation. The introduction to this special issue notes the different “monocular views” of interaction with intelligent systems. AI focused on devising better algorithms, HCI on how to improve the use of existing algorithms. AI originated in mathematics and engineering, HCI in psychology. But half a century is enough time to spawn a hybrid or synthesis had forces not pushed the two fields apart.