Designing Adjustable Work Agreements for Hybrid Intelligent Disaster Management
Keywords
Loading...
Authors
Issue Date
2021-06-13
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
This research focuses on improving first responders' situation awareness with support from an artificial intelligent agent for decisions with a value-sensitive aspect based on work agreements. To indicate the division of roles, responsibilities, and capabilities of the first responder, a team leader, and an artificial intelligent agent in the decision-making process, three Team Design Patterns were designed, evaluated, and simulated. The patterns describe how work agreements are collaboratively managed to agree on the human-AI collaboration explicitly and how these work agreements are used during disaster management to collaboratively deal with adversarial changes in weather conditions that threaten the first responder's safety. The work agreements addressed in this research indicate who to involve in the interaction, what the notification type should be, and whether task transfer from human to robot should be activated. A formal work agreement model was made to show how work agreements can be managed between more than two actors. The task allocation and information exchange between the first responder, team leader, and AI agent is simulated using the Python Rapid Experimentation software package (MATRX). In total twenty-four participants contributed to this research: eleven human-agent teaming experts from TNO (Artificial Intelligence and Human Factors researchers) and thirteen professional fire fighters from the Gezamenlijke Brandweer (five) and the Spaansland fire brigade in Enschede (eight). Results show that user-tailored notifications and behaviours of the AI agent through work agreements likely improve team performance for value-sensitive decisions significantly.
Keywords: Actionable Situation Awareness, Disaster Management, Human-AI Collaboration, Team Design Patterns, Value-Sensitive Decision-Making, Work Agreements
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen