Prince William, Duke of Cornwall, on a meet-and-greet in Tallinn, Estonia in March 2025. This image was signed by the IPTC’s C2PA WordPress Plugin at the time of publishing.
The IPTC has developed a WordPress plugin that automatically signs all images and video content published on a WordPress site. It has been put to use to automatically sign all images attached to IPTC news posts, such as this one, at the moment of publishing.
Based on our library of signing tools which are available to IPTC members, the “C2PA Signer” plugin takes action when a WordPress user publishes a new post. The plugin automatically retrieves all images (in all available sizes) and signs each image using the private key associated with the publisher.
The tool also extracts relevant metadata from WordPress. Each image’s caption, alt text, image upload date and publish date are embedded into the signed C2PA Manifest using an early version of the Origin IPTC Verified News Publisher metadata assertion. The specification of this assertion is currently in flux and the example assertion should not be relied on for production use, although the assertion is supported by the IPTC’s C2PA validator tool, Origin Verify.
This is in line with the goals of our IPTC Origin Verified News Publisher project, whereby publishers sign their own content using their own certificate. This enables publishers to take ownership of their content and to assert important facts about their content at the time of publishing.
Screenshot of the settings page for IPTC’s C2PA Signer WordPress plugin.
Screenshot of the signed Prince William photograph being viewed in the IPTC Origin Verify tool
In related news, the IPTC now has its own C2PA certificate, issued by GlobalSign under the IPTC’s official name, “Comite International des Telecommunications de Presse.” This means that the IPTC can be the first entity to use the new plugin.
“We are very happy to launch the new WordPress plugin, which we of course are using on our own website,” says Brendan Quinn, Managing Director, IPTC. “We believe that this makes us the first organisation to routinely sign all images that we publish using our C2PA credentials.”
Chinese authorities issued guidelines on Friday requiring labels on all artificial intelligence-generated content circulated online, aiming to combat the misuse of AI and the spread of false information.
The regulations, jointly issued by the Cyberspace Administration of China, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the National Radio and Television Administration, will take effect on Sept 1.
A spokesperson for the Cyberspace Administration said the move aims to “put an end to the misuse of AI generative technologies and the spread of false information.”
According to China Daily, “[t]he guidelines stipulate that content generated or synthesized using AI technologies, including texts, images, audios, videos and virtual scenes, must be labeled both visibly and invisibly” (emphasis added by IPTC). This potentially means that IPTC or another form of embedded metadata must be used, in addition to a visible watermark.
“Content identification numbers”
The article goes on to state that “[t]he guideline requires that implicit labels be added to the metadata of generated content files. These labels should include details about the content’s attributes, the service provider’s name or code, and content identification numbers.”
It is not clear from this article which particular identifiers should be used. There is currently no globally-recognised mechanism to identify individual pieces of content by identification numbers, although IPTC Photo Metadata does allow for image identifiers to be included via the Digital Image GUID property and the Video Metadata Hub Video Identifier field, which is based on Dublin Core’s generic dc:identifier property.
According to the article, “Service providers that disseminate content online must verify that the metadata of the content files contain implicit AIGC labels, and that users have declared the content as AI-generated or synthesized. Prominent labels should also be added around the content to inform users.”
Spain’s equivalent legislation on labelling AI-generated content
The Spanish proposal has been approved by the upper house of parliament but must still be approved by the lower house. The legislation will be enforced by the newly-created Spanish AI supervisory agency AESIA.
If companies do not comply with the proposed Spanish legislation, they could incur fines of up to 35 million euros ($38.2 million) or 7% of their global annual turnover.
On behalf of our memberships, IPTC and PLUS respectfully suggest that existing copyright law is sufficient to enable licensing of content to AI platforms. A “fair use” provision does not cover commercial AI training. Existing United States copyright law should be enforced.
IPTC and PLUS Photo Metadata provide a technical means for expressing the creator’s intent as to whether their creations may be used in generative AI training data sets. This takes the form of metadata embedded in image and video files. This solution, in combination with other solutions such as the Text and Data Mining Reservation Protocol, could take the place of a formal licence agreement between parties, making an opt-in approach technically feasible and scalable.
It is true that our technical solutions would also be relevant if the US government chose to implement an opt-out based approach. However, this does not currently protect owners’ rights well due to the routine activity of “metadata stripping” – removing important rights and accessibility metadata that is embedded in media files, in the misguided belief that it will improve site performance. Metadata stripping is performed by many publishers and publishing systems – often inadvertently.
As a result, we can only recommend that the US adopts an opt-in approach. We request that the US government ensures that metadata embedded in media files be declared as a core part of any technical mechanism to declare content owner’s desire for content to be included or excluded from training data sets.
Content creators are a core part of the US economy and have a strong voice. We agree with their position, but we don’t simply come with another voice of complaint: we bring a viable, ready-made technical solution that can be used today to implement true opt-in data mining permissions and reservations.
Close-up screenshot of Pinterest’s label for AI-generated content.
As reported in Social Media Today, Pinterest has started using IPTC embedded Photo Metadata to signal when content in “Image Pins” has been generated by AI.
Reports started in February that Pinterest had started labelling AI-generated images. Now it has been confirmed via an official update to Pinterest’s user documentation.
In the Pinterest documentation, a new section has recently been added that describes how it works:
Screenshot of Pinterest’s help pages showing how IPTC metadata is used to signal AI-generated content.
“Pinterest may display a label in the foreground of an image Pin when we detect that it has been generated or modified with AI. This is in accordance with IPTC standard for photo metadata. We’re working on ways to expand our capabilities to better identify GenAI content in the future through additional technologies.”
