Should Voice Agents Be Polite in an Emergency? Investigating Effects of Speech Style and Voice Tone in Emergency Simulation

Research Background and Problem

• What problems or challenges did the authors identify?
  The authors pointed out that research in human-computer interaction indicates that the politeness of voice assistants is crucial to users' satisfaction with social interactions. However, it remains unclear whether maintaining politeness is appropriate in emergency situations.

• Why is this issue important?
  In emergency situations, clear and efficient communication is critical to ensuring people's safety. The ability of voice assistants to balance politeness and directness affects their practicality as tools for issuing emergency instructions.

• Research Motivation and Related Work
  There is currently limited research on the impact of voice style (politeness) and tone (urgency) of voice assistants in emergency situations. Related fields emphasize the importance of politeness theory and perceived urgency in human-computer collaboration, particularly in influencing users' perceptions of voice assistants and their cognitive workload.

Solution

• What methods or solutions did the authors propose?
  The authors conducted an experiment manipulating the voice style of the assistant (politeness: positive politeness, negative politeness, direct) and tone (urgency: high and low) to evaluate how these factors influence user perceptions (e.g., cognitive trust, emotional trust, human-likeness, likability, intelligence) and users' cognitive workload.

• What is innovative about this solution?
  The authors systematically explored the interaction between politeness and urgency in voice assistants during emergency situations, addressing a gap in the research. Additionally, they designed a virtual experimental environment simulating emergency scenarios, using a realistic fire scene to elicit participant responses.

• What are the implementation steps and key techniques used?

  • Pre-experiments: Two pre-experiments were conducted to validate the effectiveness of the manipulations of politeness and urgency.
    • Politeness adjustment: Based on politeness theory, instructions were designed as positive politeness (e.g., "Let’s find"), negative politeness (e.g., "Can you find"), and direct commands (e.g., "Find an extinguisher").
    • Urgency adjustment: Urgency was manipulated by changing speech rate (fast vs. slow), with faster speech perceived as more urgent.
  • Main experiment: An online 3D virtual environment was constructed to simulate a fire emergency scenario, with six experimental conditions manipulated (3 levels of politeness × 2 levels of urgency).
  • Data collection and analysis: After completing the tasks, participants filled out questionnaires measuring trust in the voice assistant, human-likeness, perceived intelligence, likability, and cognitive workload.

Research Findings

• What specific findings were obtained?

  • Main effect of urgency: A high-urgency tone significantly increased users' perceptions of the voice assistant's human-likeness, likability, and intelligence, while reducing the effort and frustration required to complete tasks.
  • Politeness: Politeness did not show a significant main effect and did not notably influence users' perceptions of the voice assistant or their cognitive workload.
  • Interaction between politeness and urgency: When the voice assistant used "negative politeness" (e.g., "Can you find"), a high-urgency tone further enhanced cognitive trust and likability.

• What advantages does this solution have compared to existing ones?
  This study not only explored individual perceptions but also revealed, for the first time, the combined effects of politeness and urgency on voice assistant design. It provides empirical support for designing effective voice assistants in emergency situations.

• What were the experimental or evaluation results?
  Through statistical analysis, the authors confirmed the successful manipulation of experimental variables (urgency and politeness) and identified significant interaction effects in multivariate analysis. These findings offer guidance for designing efficient voice interactions in emergency scenarios.

• Limitations and Future Directions

  • Limitations:
    • The use of a virtual environment instead of real-life scenarios may have affected participants' perception of urgency.
    • The sample distribution (e.g., gender skewed toward female participants) and the use of only female voices may limit the generalizability of the results.
  • Future Directions:
    • Explore the performance of voice assistants in other complex tasks or different emergency scenarios.
    • Consider the influence of multicultural contexts on politeness and urgency.
    • Incorporate virtual reality technology to enhance participants' immersion and simulate real-world scenarios more effectively.

Conclusion

By systematically investigating the effects of politeness and urgency on voice assistants, this study reached the following key conclusions: urgent tones are more effective in conveying task information, enhancing trust, and reducing users' cognitive workload; under appropriate conditions, the combination of politeness and urgency can optimize user experience. These findings provide new directions for designing voice assistants in emergency situations, with potential applications in public safety, disaster management, and emergency services.