Beschrijving
Artificial Intelligence (AI) is increasingly used in military contexts as a decision support system. Many countries are developing weapons enhanced with AI capabilities, and their deployment has already started. In this context, the motivation to use AI is based on the promise of enhanced performance, including the reduction of errors during attacks. However, Human-AI interaction is not flawless and may result in errors that a lone operator would not have committed. Experimental studies on Human-AI interaction have so far primarily focused on trust in AI and design aspects influencing trust. This line of research has found that both overtrust and mistrust in AI recommendations can lead to improper human decisions and action. Nonetheless, trust does not explain the full realm of human-AI interaction failures. Therefore, the field of Human-AI interaction has drawn on studies of confidence and advice-taking in humans to deepen the understanding of operators’ behaviours. Human decision confidence and its interplaBijkomende beschrijving
Title: Decision Confidence in Human-AI interaction: A Literature ReviewPeriode | 16 okt. 2024 |
---|---|
Evenementstitel | 64TH INTERNATIONAL MILITARY TESTING ASSOCIATION CONFERENCE |
Evenementstype | Congres |
Locatie | Sydney, AustraliëToon op kaart |
Mate van erkenning | Internationaal |
Gerelateerde inhoud
-
projecten
-
Safe and Efficient Autonomous Systems for Military Operations
Project: Onderzoek