Aplicação de ferramenta open-source na validação de integridade de conteúdos digitais para o combate a deepfakes maliciosos

Abstract

The rise of generative artificial intelligence has streamlined the creation of realistic synthetic media, known as deepfakes, undermining trust in digital information ecosystems. This work proposes a framework to ensure the authenticity and provenance of digital content based on the Coalition for Content Provenance and Authenticity (C2PA) standard. The developed solution consists of a RESTful API (C2PA-API), implemented in Node.js, which automates the creation, signing, and validation of cryptographic integrity manifests. To validate the practical applicability of the proposal, a proof of concept was integrated into the AT Protocol within a decentralized social network, allowing end-users to verify media origin and editing history in real-time. Additionally, this research presents a comparative analysis between cloud infrastructure (AWS) and blockchain-based solutions, evaluating latency, costs, and energy impact parameters. The results demonstrate that the cloud approach is superior in performance and sustainability, with significantly lower energy consumption than decentralized models such as Ethereum PoS, although blockchain offers stronger guarantees of public immutability. The framework serves as a scalable tool to mitigate misinformation and restore transparency in the digital environment.