Ensuring Video Authenticity with C2PA on AWS

The proliferation of sophisticated AI-generated content tools, including Google's Veo 3 and advanced versions of Midjourney, has dramatically escalated the challenge of distinguishing authentic media from synthetic creations. As these generative technologies produce increasingly photorealistic and convincing content, the need for robust content provenance and authenticity verification has become paramount across the media industry. The Coalition for Content Provenance and Authenticity (C2PA) specification continues to serve as the leading standard for creating digitally signed metadata manifests, providing a critical foundation for content authentication in this rapidly evolving landscape. To address the growing demand for scalable provenance solutions that can handle modern streaming workflows, we present C2PA's capability in handling fragmented MP4 (fMP4) formats. Developed in collaboration with CBC, this solution extends on our previous open-source code to highlight content authentication capabilities to streaming media, enabling provenance tracking for high-quality video delivery with minimal buffering; built on AWS infrastructure using Docker and serverless technologies. As the media industry faces an unprecedented wave of AI-generated content, this scalable approach to fMP4 provenance tracking positions organizations to maintain content authenticity standards while adapting to next-generation streaming technologies. This work demonstrates practical implementation strategies for C2PA metadata, action, and ingredient assertions within fragmented MP4 streaming environments.

Published

2025-10-13

Content type

Original Research

Keywords

c2pa, fragmented mp4, content provenance, media authenticity, streaming video, aws, ai-generated content

ISBN

978-1-61482-966-9