The Algorithmic Gaze: A Multi-Modal Audit of Representational Divergence and Heritage Authenticity in Varanasi, India

Purpose: This paper introduces the Algorithmic Divergence Coefficient (Ω), the first computationally operationalised metric for measuring the statistical displacement of lived heritage authenticity by algorithmically-generated simulacra. Design/Methodology/Approach: Applying CLIP (ViT-B/32) multi-modal embeddings and cosine-distance analysis to four image corpora collected from Varanasi, India — Algorithmic (CA), Lived (CL), Platform Entrepreneurial (CEp), and Self-Represented Entrepreneurial (CEs). Findings: We derive two empirical findings: a Cultural Divergence Index (Ωc = 0.2793) confirming that AI-mediated platforms demonstrate an approximate 28% representational gap compared to the city's physical reality, and an Entrepreneurial Alignment Index (Ωe = 0.1085) revealing near-total synchronisation of local artisan visual identity with platform aesthetic templates. Social implications: Drawing on Urry and Larsen's Tourist Gaze, Baudrillard's third-order simulacra, and Latour's Actor-Network Theory, we theorise the Algorithmic Gaze as a non-human disciplinary apparatus that substitutes synthetic cultural coherence for entrepreneurially-embodied heritage authenticity. Originality/Value: In alignment with calls for middle-range theory-building and the mandate for policy-relevant social science, we propose the Omega Index as an auditable governance instrument for tourism management authorities.