Design principles for conversational agents to support Emergency Management Agencies

Area of Science:

• Information Systems
• Emergency Management
• Human-Computer Interaction

Background:

• The COVID-19 pandemic highlighted the challenge of misinformation for Emergency Management Agencies (EMAs).
• Traditional conversational agent (CA) design principles are not directly applicable to EMAs due to their unique structures and responsibilities.
• There is a need for tailored IS design approaches for CAs in emergency management contexts.

Purpose of the Study:

• To develop specific requirements and design principles for conversational agents (CAs) tailored for Emergency Management Agencies (EMAs).
• To address the challenges posed by the 'infodemic' and ensure effective public communication during crises.

Main Methods:

• Conducted semi-structured interviews with practitioners from EMAs in Germany and Australia.
• Analyzed interview data to derive meta-requirements and design principles for EMAs CAs.
• Synthesized findings to propose a basis for designing robust CAs for EMAs.

Main Results:

• Identified twelve meta-requirements for CAs in EMAs.
• Established five key design principles for developing EMAs CAs.
• Found that minimizing social cues in CAs is crucial, contrasting with traditional CA design.

Conclusions:

• The study provides a foundational framework for designing effective and robust conversational agents for Emergency Management Agencies.
• The developed principles and requirements support EMAs in disseminating trustworthy information during public health crises.
• Future IS design for EMAs CAs should prioritize functional needs over extensive social cues.