Academic journal article Human Factors

The Perceived Utility of Human and Automated Aids in a Visual Detection Task

Academic journal article Human Factors

The Perceived Utility of Human and Automated Aids in a Visual Detection Task

Article excerpt

INTRODUCTION

Information technology is changing the nature of work by providing real-time access to more relevant, accurate, and timely information than was previously possible. This has drastically increased efforts to design and develop automated aids that support, and sometimes replace, the human decision maker in military and civilian sectors (Cesar, 1995). The requirement for automation is highlighted by the increasing complexity of the military mission, which is expanding from major theater of war conflicts to include small-scale contingencies, humanitarian assistance, and peacekeeping. System developers frequently propose automated aids to enhance military decision makers' situational awareness. The underlying assumption in providing these automated aids to military personnel is that the human-computer "team" will be more productive than either the human or the automated aid would be working alone. Some researchers have found support for this underlying assumption (Corcoran, Dennett, & Carpenter, 1972; Dalal & Kasp er, 1994; Parasuraman, 1987; Thackray & Touchstone, 1989); others have found instances in which human operators err by overrelying (misuse) or underutilizing (disuse) automated systems (Parasuraman & Riley, 1997).

Automated systems must be designed and human operators trained to avoid both misuse and disuse. Cohen, Parasuraman, and Freeman (1998) blamed the widespread inappropriate use of automated decision aids as the key reason automation has not advanced as far in the decision-making arena as it has in other areas. What are the processes leading to the suboptimal performance found in human-computer teams? Because team researchers have begun to think of the human-computer system as a small group or team in which one member is not human (Bowers, Oser, Salas, & Cannon-Bowers, 1996; Bubb-Lewis & Scerbo, 1997; Scerbo, 1996; Woods, 1996), the vast literature on group dynamics and teams may be relevant to our study of human use of automated systems. Just as the human factors literature demonstrates suboptimal performance among human-computer dyads, the group dynamics literature is abundant with examples of suboptimal group performance, often referred to as process loss. Cognitive, motivational, and social processes have be en implicated in causing the process loss found in human groups (e.g., Mullen, Johnson, & Salas, 1991).

Building on this literature and Mosier and Skitka's (1996) work, Dzindolet, Beck, Pierce, and Dawe (2001) developed a framework of automation use to predict how cognitive, motivational, and social processes might affect automation reliance (see Figure 1). The purpose of the present paper was to examine one of the social processes: perceived utility, the outcome of a comparison between the human operator's perceptions of reliability of the automated aid and manual operation. In order to focus on perceived utility, cognitive and motivational processes needed to be eliminated or controlled.

CONTROLLING COGNITIVE AND MOTIVATIONAL PROCESSES

One cognitive process predicted to affect automation use is automation bias, "the tendency to use automated cues as a heuristic replacement for vigilant information seeking and processing" (Mosier & Skitka, 1996, p. 205). The fact that the automated system provides a decision may lead the decision maker to rely on this information in a heuristic manner. Conceivably, this may occur in various degrees. In its most extreme form, the decision reached by the automated aid is immediately adopted. In a less-extreme form, the decision reached by the aid may be given an inappropriately large role in the human's decision-making process. For example, Layton, Smith, and McCoy (1994) found that many pilots provided with an automated aid's poor en-route flight plan did not explore other solutions (e.g., they did not generate actual flight plans on screen) as much as pilots who were not provided with the automated aid's decision. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.