Home Cybersecurity Article

OpenAI Adds C2PA and SynthID Watermarks to AI Images

TL;DR

OpenAI implements dual authentication layers—C2PA metadata and Google's SynthID watermark—to verify AI-generated images and combat deepfakes.

Key Points

  • C2PA standard adds verifiable metadata signal to OpenAI-generated images
  • SynthID invisible watermark persists through screenshots, resizing, and digital manipulation
  • Public verification tool available to check both signals for image authenticity
  • Two-layer approach addresses individual weaknesses: metadata for information density, watermarks for durability

Why It Matters

As AI image generation becomes ubiquitous, provenance verification is critical infrastructure for detecting synthetic content at scale. This dual-layer approach sets a technical standard that security researchers and developers can build detection systems around, though effectiveness depends on industry-wide adoption beyond OpenAI's ecosystem.
Read the full announcement

Source: techcrunch.com