Easy listing, hard selling? How AI-source disclosure in second-hand product descriptions shapes purchase intention

Generative AI is increasingly used to create product descriptions in consumer-to-consumer (C2C) online second-hand markets, yet the effects of disclosing the description source remain unclear. Drawing on information diagnosticity theory, this research examines how AI- versus seller-source disclosure affects purchase intention. Across four online experiments, AI-source disclosure reduces purchase intentions in second-hand but not primary markets. This effect is sequentially mediated by perceived information diagnosticity and trust in sellers and is stronger for experience than search products. The negative effect is attenuated when AI is disclosed as being used only for proofreading or when AI-generated descriptions are reviewed by the seller. These findings identify C2C second-hand transactions as a context in which AI-source disclosure can backfire and inform the design of AI-assisted platform listing functions.