A Trust Paradox May Limit the Application of Al-Generated Knowledge

The extent of knowledge available to society may no longer be limited by human cognition or by refining algorithms. Instead, the interaction of the two may limit knowledge transfer and creation. We examine two behavioral paradoxes: Polanyi's paradox, which has existed since 1966, and the Trust paradox, which we propose. We study four technology innovations—printing, digitalization, the Internet, and artificial intelligence (AI)—and assess how these paradoxes limit the use of knowledge created by each innovation. We found that while technology has largely overcome Polanyi's paradox of limiting the transfer of tacit knowledge, the Trust paradox may limit the application of AI-generated knowledge. We identify five relevant drivers of trust—reliability, transparency, association, reciprocity, and accountability—and outline how each impacts and creates challenges for the application of AI.