When Transparency Fails: Compliance with Decision Support in Nautical Collision Avoidance
Event Type
Cognitive Engineering & Decision Making
TimeWednesday, October 12th8:45am - 9:00am EDT
DescriptionTo attempt to counter uncalibrated compliance with an automated decision aid, the current study sought to introduce some level of transparency into the system’s recommendations. An experiment was conducted using a simple, simulated maritime collision avoidance task featuring an imperfect, but highly reliable, decision aid. Specifically, the aid’s recommendations were sometimes paired with a single confidence measure associated with the actual level of difficulty of the trial. Compliance with the aid produced safer outcomes in terms of miss distance at the closest point of approach. However, transparency in the form of confidence measures did not increase overall compliance with the aid, nor affect the safety of maneuvers. Implications for transparency and compliance are discussed.