Functional Analogies Increase Trust in Black-Box AI Systems Among Lay Consumers: The Case of GeNose C-19

Abstract

Laypeople tend to distrust innovative technologies, especially when the product involves complex systems. Although researchers try to increase public trust by highlighting the accuracy of AI technologies, many people still do not trust AI products as they do not understand how the AI model generates its predictions. One solution is to help laypeople understand how AI models work using functional analogies. Functional analogies explain AI systems by building on consumers' common knowledge and drawing parallels to phenomena that consumers are already familiar with. We present a case study of a black box AI-based Covid-19 detection product, GeNose C-19, developed by the Indonesian government. We find that explaining how GeNose works using functional analogies increases both Indonesian and American lay consumers' trust in GeNose. More broadly, the findings suggest that promoters of AI products can use analogies to explain how the product works to increase lay consumers' trust in the product.ARTIFICIAL INTELLIGENCE based products often succeed in the market but often fail. Similar to traditional technological products, some market failures are due to technological shortcomings, whereas others are due to overpromises made via marketing that the technology could not deliver Computer