Understanding the Necessary Conditions of Multi-Source Trust Transfer in Artificial Intelligence

Abstract

Trust transfer is a promising perspective on prevalent discussions about trust in AI-capable technologies. However, the convergence of AI with other technologies challenges existing theoretical assumptions. First, it remains unanswered whether both trust in AI and the base technology is necessary for trust transfer. Second, a nuanced view on trust sources is needed, considering the dual role of trust. To address these issues, we examine whether trust in providers and trust in technologies are necessary trust conditions. We conducted a survey with 432 participants in the context of autonomous vehicles and applied necessary condition analysis. Our results indicate that trust in AI technology and vehicle technology are necessary sources. In contrast, only vehicle providers represent a necessary source. We contribute to research by providing a novel perspective on trust in AI, applying a promising data analysis method to reveal necessary trust sources, and consider duality of trust in trust transfer.

Towards Trustworthy AI-based Military Cyber Operations

“…Moreover, trust is an underlying function, characteristic, or value that has the power to relegate other values and is based on invisible assumptions (Simpson, 2012). In technology, trust was a central concept regarding technological acceptance and until recent efforts, it was a neglected notion in the AI domain (Renner et al, 2021). Defining and implementing this notion in technological terms is a delicate process.…”

Trust in artificial intelligence: From a Foundational Trust Framework to emerging research opportunities

“…Thus far, research on trust had predominantly focused on classes of technologies (e.g., anthropomorphic recommender agents). In prior studies, users interacted with a specific AI, with the arguments and conclusions always drawn with respect to a class of a technologies, such as recommender agents, driverless cars, virtual agents, e-vendors, robots, or chatbots (e.g., Glikson & Woolley, 2020;Lansing & Sunyaev, 2016;Renner et al, 2022). Our definition underscores the importance of examining both specific trust, as well as trust in a class of technologies.…”

“…Most research has examined the relationship in the opposite direction. Specifically, studies have examined the influence of trust in a broader sociotechnical system as a mechanism that builds trust in a particular technology that harbors the system in question (Gefen et al, 2003;Renner et al, 2022). However, with the ubiquity of AI, humans many also transfer trust from AI to the social or other technical systems within which AI is embedded.…”

“…Based on the definition of human-AI systems (Definition 3), humans, a key component of human-AI systems, are viewed broadly to include the users, potential users, policy makers, and others who interact or are potentially affected by AIbased technologies. In contrast, extant research on trust in AI predominantly focuses on humans who are in direct contact with AI; that is, AI users (Gefen et al, 2003;Glikson & Woolley, 2020;Mcknight et al, 2011;Renner et al, 2022). The latter work is of undisputed importance, as we continue to discover new facts about the nature of direct human-AI use.…”

Artificial Intelligence-Driven Convergence and its Moderating Effect on Multi-Source Trust Transfer

“…Trust transfer seems to be a promising, but so far neglected, mechanism to understand how trust may be established in converged AI-based products (Renner et al, 2021(Renner et al, , 2022. In essence, trust transfer posits that users' trust in (multiple) already existing and familiar trust sources may be transferred to an unknown target (Stewart, 2003(Stewart, , 2006.…”