TY  - CONF
AB  - When automating tasks using some form of artificial intelligence, some
 inaccuracy in the result is virtually unavoidable. In many cases, the
 user must decide whether to try the automated method again, or fix it
 themselves using the available user interface. We argue this decision
 is influenced by both perceived automation accuracy and degree of tas
k "controllability" (how easily and to what extent an automated result
 can be manually modified). This relationship between accuracy and con
trollability is investigated in a 750-participant crowdsourced experim
ent using a controlled, gamified task. With high controllability, self
-reported satisfaction remained constant even under very low accuracy 
conditions, and overall, a strong preference was observed for using ma
nual control rather than automation, despite much slower performance a
nd regardless of very poor controllability.
AU  - Roy, Quentin
AU  - Zhang, Futian (Caesar)
AU  - Vogel, Daniel
C3  - Proceedings of the 2019 CHI Conference on Human Factors in Computing S
ystems
DA  - 2019/5//
C2  - 2019
DO  - 10.1145/3290605.3300750
ID  - Roy2019_control_accu
PB  - ACM Press
SN  - 9781450359702
SP  - 1-8
TI  - Automation Accuracy Is Good, but High Controllability May Be Better
UR  - http://dl.acm.org/citation.cfm?doid=3290605.3300750
ER  -