Abstract
Artificial intelligence (AI) relies on large-scale human-produced data, yet the relationships between content creators and AI developers remain poorly understood. This study examines how these relationships are increasingly structured within AI data supply chains. We argue that emerging asymmetries in economic value distribution within these exchanges may reflect a broader pattern of exploitation by design, whereby AI developers initially position themselves as benevolent actors to facilitate large-scale data acquisition before gradually shifting toward more enclosed and extractive models. Drawing on the well-established fairness–trust connection in supply chain management literature, we conceptualise content creator–AI developer relationships as quasi buyer–supplier exchanges characterised by limited provenance, attribution, and compensation. Using an exploratory qualitative design based on interviews with content creators, AI developers, and industry experts, we develop insights into how trust may be cultivated during the early stages of exchange and subsequently undermined as value is extracted.