C2PA often dies on scheduler uploads
ImgKey, mid-2026: X strips C2PA; Graph API and cross-post paths are inconsistent; LinkedIn often preserves. Do not assume the manifest survived.
ImgKey, mid-2026: X strips C2PA; Graph API and cross-post paths are inconsistent; LinkedIn often preserves. Do not assume the manifest survived.
ImgKey, mid-2026: LinkedIn often preserves C2PA and shows a CR icon. X strips. Photoreal B2B UGC still needs a visible label.
Coinis, 16 Jul 2026: Background Generation, Image Generation, and Add Animation auto-label ads. From 1 Jun 2026 Meta also detects third-party AI.
ImgKey, mid-2026: X strips C2PA, re-encodes, and shows no credentials UI. Provenance does not travel. Disclose in the post anyway.
Skip YouTube's AI checkbox and auto-detect may still label. You can update in Studio except YouTube's own tools, verified C2PA, or T&S.
Some export paths drop Content Credentials. Know which scheduler or compressor strips provenance.
Instagram strips C2PA on upload. Re-export from CapCut often drops it. The platform then guesses.
The AI Act transparency duty is on providers and deployers. The TikTok/YouTube/Meta label is the platform's own detector. They are not substitutes.
Manifest, invisible watermark and perceptual fingerprint survive different attacks and each fails somewhere. What your provenance claim actually proves.
Meta shows AI Info on Instagram and Facebook from C2PA or IPTC signals, or from your upload disclosure. It is not a detector. A real photo can be labelled.
LinkedIn's C2PA icon reveals whether AI was used, which tool made the file, and when the credential was issued. Here is what a click shows and what strips it.
TikTok watermarks uploads that arrive with C2PA credentials attached. That makes attaching provenance a labelling decision your toolchain is already making.
A C2PA manifest is a hash over exact bytes, so re-encodes and screenshots kill it by design. How to find the hop in your pipeline where provenance is lost.
YouTube applies AI labels from three sources you don't fully control. Know which ones fire automatically before you decide what the upload toggle is for.
A Content Credential is not a watermark — it is a signed manifest of assertions about origin and edits. What is actually inside one, and what survives editing.
A Content Credential is two different bindings doing two different jobs. Walk a file through trims, re-encodes and uploads and see which one actually survives.
SynthID, a platform AI label, and a signed manifest all get called watermarking. What SynthID embeds per modality, and what it cannot tell you.