Strengthening human decision making: AI decision aids as sounding boards
Abstract
Despite the rapidly increasing use of artificial intelligence (AI) to assist decision making in many areas important to society, recent research shows that AI decision aids often do not actually improve human decision making. This proposal, at the intersection of computer science, cognitive science, and human-computer interaction, aims to radically shift the role of AI. Instead of the prevailing approach of AI suggesting and explaining recommendations, this project proposes leveraging AI as a sounding board for expert human decision-makers. Informed by recent insights from group decision making research, this inter-disciplinary project expects to yield new design guidelines, leading to more effective decision aids, and better decisions.