<![CDATA[Background: Synthetic speech and audio deepfakes increasingly unsettle legal assumptions that a recorded voice can index speaker identity, communicative intention, and evidentiary presence in Indonesian digital-public discourse. Objective: This study examines how synthetic or disputed speech becomes evidentially meaningful by analysing the linguistic authentication, evidentiary uncertainty, and procedural positioning of deepfake-related public documents in Indonesia. Method: Using a qualitative corpus-assisted legal-linguistic design, this study analyses 24 publicly verifiable documents consisting of official enforcement releases, fact-checking records, news reports, legal-context documents, and scholarly sources. Results: The findings show that authentication is constructed through convergence among detector claims, source tracing, official-account verification, platform context, and institutional classification rather than through acoustic judgement alone. Evidentiary uncertainty is not treated as a defect to be removed, but as a structured stance expressed through probability, absence, methodological caution, and legal-contextual framing. Implication: Synthetic speech is procedurally positioned most often through public verification and fact-checking before it enters, or potentially enters, formal adjudicative settings. Novelty: This study offers a novel forensic-linguistic account of deepfake evidence by shifting analysis from technical detection toward the legal-discursive process through which synthetic speech is authenticated, contested, and made evidentially intelligible under uncertainty.]]>
Copyrights © 2026
