Human-Centred Psychological Support in the Age of Artificial Intelligence: Prospects for Employee Well-Being and Organizational Trust

Abstract

Psychological support in corporations has traditionally relied on interpersonal mechanisms, including manager support, employee assistance programmes, counselling, stress-management interventions, and occupational health services. The diffusion of artificial intelligence is changing not only the instruments used to support employees, but also the psychological architecture of support itself. This study examines how corporate psychological support has evolved from predominantly human-centred models to hybrid systems that incorporate artificial intelligence, with particular attention to psychological safety, trust, autonomy, perceived organizational support, emotional disclosure, stress, burnout, and surveillance-related anxiety. An integrative conceptual review and comparative synthesis were conducted using established theories of occupational psychology and recent evidence on workplace mental health, digital mental-health interventions, conversational agents, technostress, and AI-enabled management. The analysis applies the Job Demands-Resources model, psychological safety theory, and perceived organizational support as complementary explanatory lenses. The study identifies a transition from reactive and episodic support toward preventive, continuous, and increasingly personalized support. AI can strengthen psychological resources by improving accessibility, low-threshold help-seeking, personalization, and early identification of distress. However, the same technologies can become psychological demands when employees experience monitoring as surveillance, data collection as loss of privacy, automated recommendations as reduced autonomy, or simulated empathy as depersonalized care. AI does not inherently improve or deteriorate employee well-being. Its psychological effect depends on whether it is embedded in a human-centred support system that preserves confidentiality, voluntary participation, meaningful human contact, and professional accountability. Future studies should test the proposed dual-path model across occupations and cultures and examine how human oversight, perceived organizational motives, and AI literacy moderate the relationship between AI-mediated support and employee psychological outcomes.