C2PA in DALL·E 3 will let people know if an image was generated by ChatGPT or not

OpenAI will update ChatGPT to generate C2PA metadata effectively by 12th February, according to the latest announcement made by the AI-tech company. The new capability is now available in ChatGPT for the Web that uses the DALL-E 3 technology, and by next week, ChatGPT for mobile will also generate images with C2PA metadata.

This will allow people to identify if an image was generated by ChatGPT or not, and this way it will be easier to address issues of provenance, as OpenAI states:

Images generated with ChatGPT on the web and our API serving the DALL·E 3 model, will now include C2PA metadata. This change will also roll out to all mobile users by February 12th. People can use sites like Content Credentials Verify to check if an image was generated by the underlying DALL·E 3 model through OpenAI’s tools. This should indicate the image was generated through our API or ChatGPT unless the metadata has been removed.

OpenAI

C2PA (which stands for Coalition for Content Provenance and Authenticity), is a partnership between multiple tech companies to address false information by certifying and validating media content. According to the official website, Microsoft is part of this partnership, so it makes sense that OpenAI would implement the C2PA capability within its infrastructure.

The Coalition for Content Provenance and Authenticity (C2PA) addresses the prevalence of misleading information online through the development of technical standards for certifying the source and history (or provenance) of media content. C2PA is a Joint Development Foundation project, formed through an alliance between Adobe, Arm, Intel, Microsoft and Truepic. C2PA unifies the efforts of the Adobe-led Content Authenticity Initiative (CAI) which focuses on systems to provide context and history for digital media, and Project Origin, a Microsoft- and BBC-led initiative that tackles disinformation in the digital news ecosystem.

C2PA

How to recognize if an image was generated by ChatGPT using C2PA in DALL·E 3?

Once ChatGPT serving the DALL·E 3 model generates an image, it will embed special C2PA metadata into it. This metadata will increase the size of the image, but it should be insignificant, as OpenAI says:

• 3.1MB → 3.2MB for PNG through API (3% increase)
• 287k → 302k for WebP through API (5% increase)
• 287k → 381k for WebP through ChatGPT (32% increase)

Images will then display special metadata, indicating that they were generated with ChatGPT, as you can see below.

It’s not yet known when and if OpenAI will implement this capability for older ChatGPT versions, however, for now, it will only be available on ChatGPT using the DALL-E model, which debuted in the tool last year.

What do you think about this?