Krisis Kepercayaan terhadap Citra Digital: Semiotika Visual, Deepfake, dan Negosiasi Kebenaran dalam Media Digital

Abstract

The development of generative artificial intelligence has transformed the digital visual communication ecosystem by enabling the production and manipulation of increasingly realistic images, thereby challenging the assumption that what is seen can immediately be trusted as evidence. This study aims to analyze the relationship between visual signs, deepfakes, digital verification, visual trust, and the construction of truth in digital media. This study employs an interpretive qualitative approach through visual semiotic, multimodal, and reflexive thematic analyses of a corpus of digital visual materials, including deepfakes, AI-generated images, visual manipulation, misinformation, and debates concerning authenticity. The analysis focuses on denotative and connotative aspects, the relationship between visual representation and textual framing, circulation context, and verification processes through source tracing, reverse image search, comparison with independent sources, metadata, and provenance. The findings show that visual realism still shapes initial credibility but is no longer sufficient to guarantee authenticity because deepfakes can preserve strong iconic resemblance while weakening the indexical relationship with reality. Awareness of deepfakes also creates a paradox of distrust in which both synthetic and authentic content may be questioned. Authenticity is also determined by the consistency among images, claims, sources, time, location, and circulation context. Digital verification therefore becomes an important epistemic mechanism for linking images with sources, context, and external evidence. This study concludes that visual truth is increasingly contextual, conditional, and negotiated, shifting the paradigm from seeing is believing to seeing requires verification, while emphasizing the importance of visual literacy and evidence triangulation in building trust.