DALL-E and ChatGPT Images: How to Remove the Instagram 'Made with AI' Label in 2026
Updated April 2026
Quick answer
Images generated by ChatGPT's image tool (DALL-E 3) embed C2PA Content Credentials. Instagram reads this binary certificate and applies the "Made with AI" label. To remove it before posting:
- 1Download the image from ChatGPT normally.
- 2Drop it into StripShot — the C2PA certificate will be flagged immediately.
- 3Click Strip. The C2PA container is removed at binary level.
- 4Download the clean file and upload to Instagram.
Free. Files never leave your browser. Zero quality loss.
OpenAI joined the C2PA standard in 2023 and implemented it in DALL-E. Every image you generate through ChatGPT's built-in image tool carries a signed Content Credentials certificate that explicitly identifies the image as AI-generated and names OpenAI as the creator. Instagram reads this on upload.
It doesn't matter how you edit the image afterward in ChatGPT. It doesn't matter if you crop it, add text, or adjust colors in another app before downloading. The C2PA certificate is embedded in the binary data of the downloaded file.
Which OpenAI products embed C2PA
ChatGPT image generation
GPT-4o and DALL-E 3 via ChatGPT interface
DALL-E 3 via ChatGPT Plus
All image outputs from the chat interface
OpenAI API (DALL-E 3)
Implementation varies by API version and parameters
OpenAI API (DALL-E 2)
DALL-E 2 predates C2PA implementation
Bing Image Creator (Copilot)
Microsoft Copilot uses DALL-E and embeds C2PA
What's inside the DALL-E C2PA certificate
The C2PA container in a ChatGPT image is a binary JUMBF structure (JPEG Universal Metadata Box Format) stored in APP11. It contains:
- Generator assertionIdentifies the image as AI-generated, names the model (DALL-E 3)
- Signer identityOpenAI's cryptographic certificate confirming provenance
- TimestampWhen the image was generated, signed with OpenAI's timestamp authority
- Soft bindingA hash of the image data to detect tampering
- Thumbnail assertionA compressed thumbnail of the image embedded in the certificate
Instagram checks for the generator assertion. That assertion is what causes the label. Remove the JUMBF container and the assertion is gone.
Strip it now
Remove DALL-E / ChatGPT metadata. One click.
JPEG, PNG, WebP · Up to 25MB · 3 free strips/day
Drop images here or browse
Select files· JPEG, PNG, WebP · Up to 20 at once
Step-by-step: clean your ChatGPT image before posting
Download your image from ChatGPT
Right-click the image in the ChatGPT interface and save it. The downloaded file contains the C2PA certificate.
Drop it into StripShot
StripShot scans the file. You'll see a red finding: 'Instagram Will Flag This as AI' with the C2PA container detected.
Click Strip and Download
One click removes the C2PA certificate, OpenAI XMP creator fields, and all other metadata. Binary-level. Zero quality loss.
Post to Instagram
No C2PA data. No OpenAI signature. Instagram reads nothing. The 'Made with AI' label doesn't appear.
Frequently asked questions
Do ChatGPT images embed C2PA metadata?
Yes. Images generated through ChatGPT using DALL-E 3 embed C2PA Content Credentials. OpenAI joined the C2PA standard and implemented it in their image generation pipeline. When you download an image from ChatGPT's image creation tool, it contains a binary certificate identifying it as AI-generated.
Does the DALL-E API embed C2PA?
The DALL-E API (direct API calls) may not embed C2PA in all cases, as the implementation can vary by API version and request parameters. Images generated through ChatGPT's built-in image tool are more consistently signed with C2PA.
Will OpenAI images always trigger the Instagram AI label?
Images downloaded from ChatGPT's image generator reliably trigger the label due to C2PA. API-generated images may or may not, depending on how the image was generated and saved. When in doubt, run the image through StripShot to check before posting.
Does removing the C2PA certificate change the image quality?
No. C2PA is stored in a separate metadata container, not the image pixel data. Binary-level removal drops the container without touching the image data. Pixels are bitwise identical before and after stripping.
Is the C2PA certificate the same structure as in Adobe Firefly images?
Yes. C2PA is an open standard, so the certificate structure is identical whether it comes from ChatGPT/DALL-E, Adobe Firefly, or Microsoft Copilot. The signer is different (OpenAI vs Adobe vs Microsoft), but the container format is the same. StripShot removes C2PA regardless of signer.
What other metadata do DALL-E images contain?
Beyond C2PA, DALL-E images typically contain XMP metadata identifying OpenAI as the creator, a generation timestamp, and sometimes model version information. StripShot removes all of this in a single pass.
Related guides
Remove the Made with AI Label
Adobe Firefly
Firefly Instagram AI Label
Midjourney
Midjourney Instagram Label
DALL-E
Remove All DALL-E Metadata
Clean your ChatGPT images before posting
Strip DALL-E C2PA. Free.
3 free strips to try it. No credit card. Pro is $9/month for unlimited.