Learn
Straight answers about AI content marking.
The field is full of confident claims that don't survive scrutiny. These explainers stick to what is actually true — about watermarks, C2PA, and what the law requires.
8 min read
AI watermarks explained: what they prove and what they don't
The two different mechanisms behind AI content marking, why people confuse them, and the one distinction that prevents false accusations.
Read7 min read
Which platforms strip content credentials?
Original research. Instagram and WhatsApp destroyed a signed file's credential and its AI declaration. Gmail and Google Drive preserved both. Measured, not assumed.
Read6 min read
Does Claude watermark its text?
Yes, since 2 August 2026, everywhere in the world. Why nobody outside Anthropic can read it yet, and why a detected mark would not prove what people expect.
Read6 min read
Does ChatGPT watermark its text?
No — OpenAI built the technology and shelved it. But its images carry credentials you can verify today, and the hidden characters people keep finding are not what they think.
Read7 min read
What is SynthID?
Google's watermark lives inside the pixels rather than the metadata, which makes it durable — and means only Google can currently read it.
ReadMore explainers coming
Deep dives on C2PA verification, metadata forensics, and provenance for procurement are in the works.