Abstract

Gambling disorder and gaming disorder are now both recognized behavioral addictions, in the DSM-5 and ICD-11 respectively, and their products are converging: loot boxes, gacha, social casino games, and skins and esports betting blur the line between playing and wagering. The empirical literature finds these gambling-like features reliably but moderately associated with gambling-related harm, while no causal link has been established. That associational, non-causal evidence raises the conceptual problem this paper takes as its starting point: how should responsibility be conceived when a designed, commercial product is reliably associated with harm yet cannot be shown to cause it? Its answer is the adaptive machine zone, a state of sustained engagement in which a monetized system uses behavioral data about the user to personalize offers, pacing, friction and prompts so as to narrow the opportunity to disengage. The original contribution is not to identify gambling-like mechanics but to specify behavioral personalization as the threshold at which responsibility shifts from individual self-control to accountable adaptive design. The paper argues that agency and responsibility are, to that extent, distributed across the human-machine system, and that a commercial-determinants framing can license precautionary, design-level regulation without individual-level causation. The framework is deliberately bounded, applying only where personalized, adaptive design meets concentrated harm. A proof-of-concept audit of ten high-grossing titles surfaces an observability gradient: the precondition features are publicly verifiable, but the adaptive features that define the framework are not, so that current transparency regimes make the relevant question unanswerable from outside.