tech
January 21, 2026
Sora is showing us how broken deepfake detection is
Posts from this topic will be added to your daily email digest and your homepage feed.
TL;DR
- OpenAI's Sora 2 can generate highly realistic deepfake videos, including those of famous people and copyrighted characters.
- The C2PA Content Credentials system, designed to authenticate digital media, is not being effectively implemented or clearly displayed on most platforms.
- Many social media platforms, including Instagram, TikTok, and YouTube, have barely visible or easily missed labels for AI-generated content.
- Metadata used for C2PA can be stripped by platforms, and current detection methods often require significant user effort.
- Industry experts believe that a combination of C2PA, inference-based AI detection tools, and legislative action is necessary to address the deepfake problem.
- Adobe is advocating for legislative solutions, such as the FAIR Act and PADRA, to protect creators from AI impersonation.
- There is a reliance on the good graces of tech companies to self-police, which is proving insufficient.