Reconfiguring Authenticity in the Age of Artificial Intelligence: A Critical Code Analysis of Deepfake Generation and Detection Systems
Abstract This study examines how authenticity is constructed and reconfigured within deepfake generation and detection systems. Drawing on the perspective of Science and Technology Studies (STS) and the methodological framework of Critical Technical Practice (CTP), it conducts a critical analysis of the core code of two representative open-source projects, FaceFusion and DeepfakeBench. Through the syntactic, semantic, and pragmatic dimensions of code, the study investigates how authenticity judgments are technically operationalized within deepfake systems. The findings reveal three interrelated processes through which authenticity is reconfigured in deepfake technological systems. First, authenticity undergoes Proxyization, as direct judgment of authenticity itself is increasingly suspended and verisimilitude comes to function as its technological proxy. Second, authenticity undergoes Riskification, as authenticity judgment shifts from factual verification to risk calculation and risk balancing. Third, authenticity undergoes Residualization, as technological systems establish authenticity by continuously identifying and excluding falsity. Consequently, authenticity no longer functions as a positive category requiring direct confirmation. Instead, it becomes a negative category established through the negation of falsity and exists only as the residual state left after falsity has been continuously identified and excluded. More importantly, these processes demonstrate that deepfake technologies are driving a profound transformation of the digital information environment: credibility is increasingly established through the exclusion of falsity rather than the confirmation of truth, thereby reshaping the fundamental logic through which authenticity is established and trust is sustained within the broader social information environment.
Authors
- Ruixin Hu
Institutions
- Xi'an Jiaotong University (CN)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-17
- DOI
- https://doi.org/10.21203/rs.3.rs-10519755/v1
- Primary Topic
- Misinformation and Its Impacts
- Type
- preprint