AI labelling in the content process: InterRed makes AI involvement traceable across all channels

Transparency requirements for AI-generated content

The use of AI is an integral part of modern content production. Article 50 stipulates transparency and labelling requirements for certain artificially generated or modified image, audio and video content. AI-generated or AI-modified texts may also be subject to labelling requirements, for example when they provide information on topics of public interest. There is no blanket labelling requirement for all AI-assisted content: the decisive factors are the type of content, the extent of the modification, the purpose of publication and any editorial review. The formal approval of an AI output does not replace editorial control over the content.

Reliably recording AI use in the CMS

To ensure transparency, information regarding the use of AI must be recorded at the time of creation, editing or import of content and must not be lost when the content is subsequently used.

To this end, InterRed has developed a technology for persistent AI tagging: information on the use and nature of AI involvement is technically linked to the respective content and retained throughout the entire publication process. The labelling is thus retained even during preparation, conversion and delivery in various formats and channels – from websites, apps and e-papers right through to printed products. This ensures consistent, cross-channel traceability of AI involvement.

For images, InterRed can automatically detect whether an asset has been generated or modified using AI from one of the major providers. Origin information and machine-readable tags are retained throughout the subsequent content process.

With text, too, InterRed enables editorial teams to trace whether and in what form AI has been involved. Editors can review AI-generated content, trace the editing steps and consciously assume editorial responsibility. InterRed thus supports a clear distinction between technical AI support, human review and final publication.

InterfaceAI integrates AI functions with editorial workflows

With InterfaceAI, customers can utilise AI functions in a similar way to MVC, but integrate any AI models directly on the client side. Secure labelling is also carried out here. The information captured in this process can be integrated into InterRed’s overarching content process and tracked consistently across all output channels.

InterRed consolidates information on the creation and editing of content within a seamless, AI-supported multi-channel publishing process. Content is created, reviewed and approved centrally, and then distributed to websites, apps, social media, newsletters, e-papers and print media. Relevant details regarding AI involvement are retained even when the content is reused or adapted for specific channels.

For publishers, media companies and organisations with extensive publication processes in particular, InterRed thus combines transparency with efficiency: AI interventions in texts remain visible, the AI status of images can be retained and carried forward through subsequent processing steps, and editorial review and approval processes can be seamlessly integrated into existing workflows.

"AI support is now part of everyday life for editorial departments and content teams. This makes it all the more important that not only the final result, but also its creation process, remains traceable. InterRed makes AI contributions within the content process visible. With SmartAI and InterfaceAI, InterRed combines this transparency with efficient web, print and multi-channel workflows," says Prof. Dr André Klahold, CEO of InterRed.