Algorithmic Provenance: When and Why Consumers Discount Products Made With AI ‐Involved Processes

Firms increasingly incorporate generative and non‐generative artificial intelligence (AI) into product design, formulation, and production, and they face a disclosure question: how should they describe the role of AI to consumers? Across six studies, including analysis of 1800 marketplace listings and five pre‐registered experiments (total experimental N = 2760), we introduce and test a 2‐by‐2 framework of algorithmic provenance defined by the AI's role (agentic vs. assistive) and its visibility in the final product (visible vs. invisible). Identical products are evaluated differently depending on where they fall in this taxonomy. Agentic AI involvement (AI as designer) reduces consumer product attitudes, perceived quality, purchase intentions, and willingness to pay relative to assistive AI involvement (AI as a tool for a human designer). The effect is amplified when AI involvement is visible in the final product. The downstream responses are partially mediated by a new construct, perceived authenticity of creative intention, which captures consumers' inference about whether a human creator pursued a genuine creative vision. The effect is larger for expressive than for functional product domains and larger among consumers primed to hold an independent self‐construal. A scalable managerial intervention, creative‐intent framing, narrows the agentic penalty by approximately half, again through the authenticity mechanism. The findings identify a substantively important new construct for AI disclosure research, refine theories of algorithmic aversion, and offer marketers an actionable disclosure design principle.