AI Watermark Center
OpenAI research

OpenAI Image Watermark

Supported OpenAI-generated images use layered provenance: SynthID, Content Credentials, and optional visible disclosure. Coverage varies by product, date, and export path.

Published
Last verified
Review
Reviewed against primary sources

Supported OpenAI-generated images use layered provenance rather than one single visible watermark. A visible OpenAI disclosure, an invisible SynthID watermark, C2PA Content Credentials, and an OpenAI verification result are four different things. Do not collapse them into “the ChatGPT watermark.”

Separate provenance layers
Supported OpenAI-generated images can carry SynthID, C2PA Content Credentials, and an optional visible disclosureA supported OpenAI image output splits into three independent branches: optional visible disclosure if requested in a prompt, invisible SynthID, and C2PA Content Credentials. An OpenAI verification result is a separate provider check.Supported OpenAI image outputVisible disclosureprompt-requestednot a default overlaySynthIDinvisiblenot AWC-detectableC2PAContent CredentialsJPEG/PNG/WebP inspectable

Visible disclosure, SynthID, and Content Credentials are independent. An OpenAI Verify result is a fourth concept: a provider check, not a watermark type.

1. SynthID

SynthID is an invisible watermark embedded in supported generated media. Users cannot see it. OpenAI documents it on supported images from ChatGPT, Codex, and the OpenAI API, complementary to C2PA. If metadata is removed, OpenAI says the watermark may still provide a signal. AI Watermark Center does not currently detect SynthID.

2. C2PA / Content Credentials

Content Credentials are C2PA provenance metadata. On supported OpenAI images they can travel with the file and may include origin information such as issuer, model, and generation time when present. They can be stripped by screenshots, some conversions, and some re-exports. This site can inspect embedded credentials in JPEG, PNG, and WebP. Finding generic C2PA is not the same as finding an OpenAI-signed manifest.

3. Visible disclosure

OpenAI documents that a visible disclosure can be requested in a prompt on ChatGPT, Codex, and OpenAI API images. That is not a default logo overlay on every generated image, and it is not SynthID or C2PA. Absence of a visible graphic is not “no SynthID” and is not “no C2PA.”

4. Public OpenAI verification

OpenAI’s official verifier can look for supported OpenAI C2PA and SynthID signals on supported image files (and SynthID on supported audio). A detected signal is described as indicating generation by, or export from, OpenAI tools. It is not a guarantee of accuracy, ownership, or unedited content. That check is a provider result, not an AI Watermark Center detection.

Official OpenAI verification (opens in a new tab)External provider tool

5. What AI Watermark Center can inspect

  • Embedded C2PA in JPEG, PNG, and WebP: available.
  • Ordinary EXIF, XMP, IPTC, and GPS when present: available.
  • SynthID: not available.
  • OpenAI Verify / Content Provenance API: not integrated.

Inspect Content Credentials

Current limitations

Current Help Center documentation covers supported images generated with ChatGPT, Codex, and the OpenAI API. Coverage may vary by product, model, export path, file type, and generation date. Do not assume every historical OpenAI image contains both C2PA and SynthID.

This pillar explains OpenAI’s documented image-provenance architecture. ChatGPT-specific user questions belong on the ChatGPT image page. It is not a substitute for the image-watermark technology overview, and it is not a SynthID detector.

Sources

  1. Provenance signals (Content Credentials, SynthID) in OpenAI-generated content — OpenAI Help Center

    Accessed August 17, 2026.

    Primary source for current OpenAI image C2PA + SynthID on supported ChatGPT, Codex, and API outputs; audio SynthID; prompt-requested visible disclosure; and openai.com/verify. Visible marks are separate from embedded signals.

  2. Advancing content provenance for a safer, more transparent AI ecosystem — OpenAI

    Published July 31, 2026. Accessed August 17, 2026.

    31 July 2026 update: supported audio SynthID, audio on the public verifier, and verification API access. Historical context for layered C2PA + SynthID; current product scope follows the Help Center.

  3. Verify — OpenAI

    Accessed August 17, 2026.

    Public verification surface referenced by OpenAI Help Center for supported image and audio provenance signals (C2PA and/or SynthID). External provider tool — not an AI Watermark Center feature.

  4. Content provenance — OpenAI API documentation

    Accessed August 17, 2026.

    Content Provenance API (also referenced as OPENAI-API-PROV-001 in Phase 7 notes). Images: C2PA + SynthID. Audio: SynthID. Documents not_detected semantics. Public/documented; not integrated by AI Watermark Center.

In this cluster