When the Vault is Never Opened: AI‐Enabled Inference and the Future of Consumer Privacy

Abstract Most research on consumer privacy assumes that privacy risks stem from unauthorized access to or misuse of sensitive consumer data. This review highlights the need to reassess consumer privacy in response to two ways in which artificial intelligence (AI) is transforming consumer privacy risk. First, AI widens the gap between what consumers intentionally reveal and what can be inferred about sensitive attributes from seemingly nonsensitive information. This increases the risk that privacy harm may arise even when sensitive data were never requested, disclosed, or accessed but instead inferred. Second, AI expands inferential capacity beyond firms to other consumers and non‐firm actors, increasing the risk of interpersonal privacy violations. Building on these arguments, I illustrate five types of AI‐enabled privacy harm and show how each can arise across organizational and interpersonal contexts. Given consumers' limited ability to anticipate, observe, or control AI‐enabled inferences or their consequences, these emerging risks create new challenges for privacy theory, management, and regulation while raising important questions for future research on consumer privacy.