Short Title:
Conversation Trust Analysis
Responsible HRP Element:
Human Factors and Behavioral Performance
Funding Status:
Active - Currently funded and in progress
Procurement Mechanism(s):
|
|
Aims:
This task is part of the Human Capabilities Assessments for Autonomous Missions (HCAAM) Virtual NASA Specialized Center of Research (VNSCOR) deliverable. This task aims to create conversational measures and mechanisms that allow virtual agents to use an unobtrusive approach to determine if a user is over/under trusting and mitigate this miss calibration with corrective actions and indications within the conversation.
- Aim 1: Develop real-time conversational indicators of trust.
- TOCA aid (Total Organic Carbon Analyze)—As a prototypical virtual assistant
- Develop conversational indicators of trust from the AMO TOCA study data
- Develop and validate TOCA as a microworld testbed
- Aim 2: Assess robustness of real-time conversational indicators of trust under long duration mission stresses.
- Data collected in a manner that mirrors the experiment from Aim 1 that used the TOCA microworld
- Aim 3: Develop and evaluate conversational affordances for managing trust. Similar to Aim 1, this evaluation will be a ground-based study using participants from University of Wisconsin-Madison and NASA JSC who will work with an intelligent virtual agent in the TOCA microworld.
- Validation in the HERA environment
- Develop principles and guidelines for trust measurement and calibration
Ground Analog Resources
Ground-Based Flight Analogs
- Controlled Research Analog, Short Duration (2 months or less) – isolated and confined with spaceflight mission scenario [e.g., HERA, NEEMO]
|
Number of Subjects
|
16
|
Category:
Risk Characterization, Quantification
Subcategory:
Evidence or Risk Characterization
Description:
The outcomes of this research will promote more effective interactions and acceptance of virtual assistants, and will provide new analytic techniques for understanding how people work with automated agents as team members. Three advances for trustable technology require the development of new analytic techniques for understanding human interaction with automated teammates.
Deliverables include:
- Interim definition phase progress report
- Develop and document principles and guidelines for trust measurement and calibration
- Paper and tool for trust estimation
- TOCA testbed
- Final report with guidelines and standards on trust and technology
Internal Customers:
Human Factors and Behavioral Performance
External Customers:
None