Related Experiment Videos
Automation bias and errors: are crews better than individuals?
L J Skitka1, K L Mosier, M Burdick
1Department of Psychology, University of Illinois at Chicago, USA. lskitka@uic.edu
The International Journal of Aviation Psychology
|September 7, 2001
Summary
Automation bias, the tendency to over-rely on automated systems, persists even in teams. Training reduced some errors, but both teams and individuals still struggled with automation failures and conflicting information.
Area of Science:
- Human-Computer Interaction
- Cognitive Psychology
- System Safety
Background:
- Automated decision aids can lead to reduced vigilance, a phenomenon known as automation bias.
- This bias may stem from the human tendency to exert less cognitive effort when automation is available.
Purpose of the Study:
- To investigate whether automation bias is mitigated in two-person crews compared to solo operators.
- To examine the impact of specific training on automation bias in different team structures.
Main Methods:
- A study comparing automation bias in solo performers versus two-person crews.
- Varying instruction conditions were applied, and training focused on automation bias and errors.
- Performance was assessed based on responses to system irregularities and automated directives.
Main Results:
- Training successfully reduced commission errors but did not significantly impact omission errors.
- Both teams and solo performers exhibited similar levels of automation bias.
- Participants failed to respond to unindicated system events and incorrectly followed automation when it contradicted other information.
Conclusions:
- Automation bias remains a significant concern, even within a team context.
- Current training methods may not fully address all facets of automation bias, particularly omission errors.
- Further research is needed to develop effective strategies for mitigating automation bias in complex systems.