Agent Transparency, Situation Awareness, Mental Workload, and Operator Performance: A Systematic Literature Review

Area of Science:

• Human-Computer Interaction
• Cognitive Systems Engineering
• Automation and Control

Background:

• Sophisticated automation in safety-critical domains necessitates effective human oversight.
• Automation transparency, enabling "seeing-into" an agent's processes, is a potential design principle to aid human supervision.
• Existing empirical support for automation transparency's impact on human factors is not systematically documented.

Approach:

• A systematic literature search adhering to the PRISMA method was conducted.
• The search focused on empirical research examining transparency's effects on key human factors variables.
• Seventeen experimental studies investigating transparency in controlled human-automation interaction settings were included.

Key Points:

• A general trend suggests beneficial effects of transparency on human performance.
• Variations in Situation Awareness, mental workload, and operator performance were observed.
• Effects are task-dependent, influenced by agent type and transparency information integration.

Conclusions:

• Automation transparency shows promise for improving Situation Awareness and operator performance.
• Benefits are evident in scenarios involving responses to agent proposals and supervisory control.
• Increased mental workload is not a necessary cost associated with enhanced transparency.