Is an image AI-generated? Detection, labeling & the law
How to check, and what the rules (EU AI Act) require
As AI images become hard to distinguish from photos, two questions matter: how can you tell if an image is AI-generated, and what do the rules require for labeling it? This guide covers both — and how machine-readable marking like C2PA fits in.
How to tell if an image is AI-generated
There is no perfect detector, but you can look for signals: AI tools often embed generation metadata (prompt, model, settings) in the file, and a growing number attach C2PA Content Credentials. The ExifGhost viewer reads both. Treat these as evidence, not proof — metadata can be stripped or edited, so absence of a signal doesn’t guarantee an image is authentic.
The EU AI Act
The EU AI Act (Regulation (EU) 2024/1689) is the first comprehensive law on AI. Its transparency rules (Article 50) require providers of generative-AI systems to mark synthetic image, audio and video output in a machine-readable way so it can be detected as artificially generated, and those who deploy systems producing “deep fakes” to disclose it. These obligations become applicable in 2026.
Who is responsible?
The obligations fall on the providers and deployers of the AI system that creates or manipulates the content — not on a neutral metadata tool. ExifGhost helps you read provenance and, if you choose, add a machine-readable C2PA marking; using it does not by itself make you compliant.