Enhancing Decision Making through Adaptive Trustworthiness Cues

ABSTRACT: Gaining tactical advantage, ensuring safety, and completing missions are the main objectives of military operations. Accomplishing these objectives depends on the ability of military personnel to make quick, effective decisions. Systems must be developed that will provide users with assistance in quickly attributing an appropriate level of trust to individual data points as well as to synthesized data, thereby supporting users in quickly developing an accurate understanding of the situation and in making decisions that effectively support mission goals. This effort will utilize a multi-faceted approach called"Adaptive Calibration of Trust for Automated System Technologies (ACTFAST)"to address this challenge. ACTFAST will be comprised of the following components: (1) a mapping of the key factors of trust to the operator"s information requirements, (2) prototype design concepts for visualizing appropriate levels of trust in information, (3) adaptive interface strategies to aid operators in appropriately calibrating trust, (4) metrics for evaluating the level of human+machine trust, (5) metrics for assessing the effectiveness of the resultant ACTFAST model to enhance decision making, and (6) metrics to assess operator performance given their trust level and the effectiveness of ACTFAST. BENEFIT: Appropriate calibration of trust in system data and system recommendations is critical for efficient, effective decision making. The development of an adaptive GUI for supporting trust calibration across a variety of operational, system, and operator states would be a valuable contribution for any government or commercial organization that relies on automation for performance of critical functions. ACTFAST has commercial potential for any number of complex, operational environments that require operators or analysts to develop and/or evaluate trust as an element of their interaction with automation. Properly calibrated trust in automated systems leads to better situation awareness, which in turn leads to better decision making. Users of all automated systems need accurate and adaptable support for trust in the system and in the data it provides. Users that would benefit from an adaptable trust-support system include unmanned vehicle operators, cybersecurity analysts evaluating attack signatures, oil pipeline control room operators, and a variety of other domains in which the trustworthiness of the automation and the data from it must be assessed as part of developing one"s situation awareness. As such, ACTFAST is expected to appeal to both other government organizations and private sector targets. The majority of the development costs will be covered through the SBIR program, leaving acceptable levels of risk for any costs that must be assumed to modify and demonstrate the capability of ACTFAST in a new domain.