AI Tool of the Week: This OpenAI tool tells you if an image is AI-generated

1 hour ago 1

ARTICLE AD BOX

logo

Discover OpenAI Verify, the groundbreaking tool that allows you to confirm the authenticity of images before sharing them.(Pexel)

Summary

A free OpenAI tool now lets users verify whether an image was generated by AI, using hidden watermarks and metadata that survive screenshots, edits, and format changes.

The AI tool we unlocked today is: OpenAI Verify.

What problem does it solve?

Here is a friction point that is becoming impossible to ignore: you receive a photograph, product shot, or campaign image, and you have no reliable way to know whether it was AI-generated. The image looks real, and you are expected to make a judgment call with nothing but instinct.

OpenAI Verify addresses this directly.

Upload any image, and the tool checks for two provenance signals: C2PA Content Credentials — cryptographic metadata that identifies origin — and SynthID, an invisible watermark embedded inside the image pixels that survives screenshots, crops, and format conversions.

A positive match confirms the image came from ChatGPT, the OpenAI API, or Codex, along with details on how and when it was created.

Free, publicly available in research preview, and no account required.

What can it do?

Confirm before you share: check any image before publishing, forwarding, or approving it for official use.

Surface what metadata hides: SynthID detection works even after screenshots or edits strip away the original file metadata.

Build a verification habit: a 30-second drag-and-drop check that turns provenance from a vague concern into a concrete workflow step.

Example

A communications lead at a financial services firm receives campaign images from an external agency. Before sign-off, she wants to confirm none were AI-generated without disclosure.

Upload the image: drag and drop or click to upload one image at a time.
Run the check: the tool scans automatically on upload. No prompt required.
Read the result: It returns whether it detected C2PA metadata, a SynthID watermark, or no supported signal.
Act on the finding: a positive signal confirms OpenAI origin along with creation details. A negative result narrows, but does not eliminate, the possibility of AI involvement.
Switch formats if needed: If the original file returns an error, re-uploading as a cropped screenshot can improve signal detection.
Document the result: screenshot the verification outcome for internal records or compliance reporting.
Edge case worth knowing: a missing signal does not mean the image is not AI-generated. Metadata can be stripped, watermarks can degrade, and the tool currently detects only images generated using OpenAI models, not other AI platforms.

What makes OpenAI Verify special?

Dual-signal detection: It combines C2PA metadata and SynthID watermarking, catching what either method alone might miss.

Watermarks that survive edits: SynthID is embedded inside image pixels, not file headers, allowing it to persist through screenshots, resizing, and format conversions.

Free and public: anyone can use it to verify content, not just OpenAI customers or paying users.

Mint's ‘AI tool of the week’ is excerpted from Leslie D'Monte's weekly TechTalk newsletter. Subscribe to Mint's newsletters to get them directly in your email inbox.

Note: The tools and analysis featured in this section demonstrated clear value based on our internal testing. Our recommendations are entirely independent and not influenced by the tool creators.

Jaspreet Bindra is co-founder and CEO, and Anuj Magazine is co-founder, of AI&Beyond.