As social media platforms are increasingly flooded with AI-generated visual content, further blurring the lines between recorded footage and virtual worlds, previous understandings of realism and visual truth are being renegotiated along the lines of platform economies and politics. New forms of digital visual literacy are required to navigate these emergent synthetic realities. The proposed presentation will outline theoretical approaches to digital visual literacy in the context of synthetic visual worlds and engage in a critical discussion of how they can be operationalized in the analysis of references to religion(s) within AI-generated visual material. The proposed presentation will draw on examples from the research project «Synthetic Realities» which focuses on US-American and European right-wing uses of AI-generated visual material on social media and the different roles visual references to religions take on therein. As recent examples, such as the current US government's social media accounts, have illustrated, generative AI lends itself well to reactionary causes, not least due to its infrastructurally conditioned tendency to reproduce stereotypes. In the context of religion, this implies that religion is often visible as a marker of identity or otherness, thereby reifying inequalities. Fostering an understanding and critical reflection of these representational tendencies and the platform and infrastructure logics that underpin them is a central concern of digital visual literacy in the age of generative AI. Therefore, a notion of digital visual literacy that combines representation-focused and platform-critical competencies is required. Drawing on a combination of concepts from critical visual literacy, platform literacy, and critical AI literacy, the proposed presentation aims to sketch different dimensions of skills and competencies that are needed for users to navigate these emerging synthetic realities.