Sora-2 is a nightmare. But there’s an upside to the visibility it’s bringing to AI-generated content.
The popularity of OpenAI’s “sloptok” AI-video generation and sharing platform is precipitating something that was going to happen anyways. Very quickly, the perception of content online is shifting from default real to default fake.
With this comes the realization that authenticating human-created content is now more important than “detecting” content which is AI-generated. It never worked well anyways and created many secondary problems.
People are finding out about the effort behind Coalition for Content Provenance and Authenticity (C2PA). The solutions that provide authentication are vastly different from those that performed detection. They depend on the content, the workflow that created the content. They also rely on fundamental software engineering principles such as cryptographic manifests, digital signatures and structured metadata embedding - the kind of stuff that can’t be vibe-coded.