The Privacy–Support Paradox of AI-Based Employee Stress Management: The Role of Perceived Surveillance, Explainability, Trust, and Employee

Abstract

Abstract New organizational ATPs are adopting artificial intelligence (AI) across various tools to detect and mitigate employee stress – sentiment analysis, facial-emotion analytics, wearable biosensors, and AI-powered counselling via chatbots. While the technology behind these systems should make effective support for wellbeing more available, and therefore earlier intervention, it also should make it possible to monitor wellbeing (The Conversation 2025). This conceptual paper brings together the results published between 2015-2026, in order to explain the ‘privacy–support paradox' which is the tracing of perceived supportive stress management technology losing some of its supportiveness due to its perceived intrusiveness of the monitoring this entails. Theoretically, the paper borrows concepts from the privacy calculus theory (Laufer & Wolfe, 1977), the Antecedents–Privacy Concerns–Outcomes (APCO) framework (Smith, Dinev, & Xu, 2011), organizational surveillance theory, theories of trust in automation, and self-determination theory (Deci & Ryan, 1985; Ryan & Deci, 2000), to propose a conceptual model where the perceived surveillance is the primary psychological mechanism that suppresses employee engagement with wellbeing AI tools. In this model, explainability, trust, and consent become boundary conditions to reduce (but not eliminate), the paradox. The paper argues an agenda of nine research propositions, offers theoretical contributions and provides managerial implications and a research agenda for the future.