Abstract
Unreliable automation may be usefully applied to situation assessment tasks, but it requires careful analysis of task demands and good interface design. Here we analytically and empirically investigated the case of target detection in which more and less reliable automation identified promising search locations to guide users' attention. Participants searched a field of 44 locations to find four targets. Automation more or less reliably marked some locations as more promising than others. Participants could use even the less reliable automation to direct and facilitated their search compared with manual undirected search. A richer automated marking scheme proved useful but required a more complex search strategy that sometimes hindered participants.
Get full access to this article
View all access options for this article.