Rethinking artificial intelligence in digital mental health through innovation illusion and inequality in the Global South

Artificial intelligence (AI) is increasingly promoted as a transformative response to the global mental health crisis, particularly in contexts where professional services remain scarce and unevenly distributed. Yet the dominant innovation narrative often overlooks the relational, epistemic, and structural conditions that shape whether AI can produce meaningful and equitable mental health outcomes. This conceptual perspective reconceptualizes AI in digital mental health through an integrative framework of innovation, illusion, and inequality. Rather than offering a systematic, scoping, or meta-analytic review, this article develops a theory-building synthesis that integrates literature on AI-enabled service expansion, artificial empathy, therapeutic alliance, algorithmic bias, and Global South inequalities. The framework argues that AI may expand access and scalability, but its effectiveness is constrained by the illusion of empathy, the limits of computational understanding, and the unequal distribution of digital infrastructure, data representation, and governance capacity. Drawing on socio-technical and health equity perspectives, the article positions AI not as a neutral clinical tool but as a system embedded within global power asymmetries. The contribution of this article lies in synthesizing fragmented debates into a context-sensitive conceptual model that explains how innovation, illusion, and inequality interact to shape user engagement, therapeutic quality, and mental health outcomes. This perspective calls for AI implementation that is clinically cautious, culturally responsive, ethically governed, and oriented toward equity rather than efficiency alone. Not applicable.