Per Erséus
Since 2 August 2026, the EU AI Act has required that AI-generated and AI-manipulated content be clearly labelled. This applies above all to so-called deepfakes: images, audio and video created or altered with AI in a way that makes them appear authentic. It doesn't have to involve people; it could just as easily be a place, an environment or an event that looks real to the viewer.
If you work with large volumes of images and video, the question is a practical one: how do you keep track of what's AI, and how do you make sure that information follows the file all the way out to where it's published? That's exactly what Mediaflow's new AI Content Labeling feature solves.
You mark a file as created with AI or edited with AI, and can add a short free-text note about how AI was used, what was changed, or where the file comes from. The label then travels with the file everywhere it appears in Mediaflow, in lists and previews, and it's available through our API and in the Photoshop and InDesign plugins.
A few things that make it easy in everyday work:
Setting, changing or removing a label requires permission to index files. Support for portals and the Office integration will come in a later release.
Until now, many users have handled AI labelling with colour markings and custom fields. That works too, but it assumes everyone knows what, say, red or orange means. AI Content Labeling instead gives you a clear icon and a free text per file, understandable even for web editors and others outside administration. Under article 50 of the regulation, no particular symbol is required, but the disclosure must be clear and distinguishable at the point of first interaction, and it must be evident where the content comes from. So text can be enough. Another advantage is that your colour markings are freed up for other uses.
Keeping pace with new legislation is nothing new for us at Mediaflow; we've previously built support for both GDPR and the accessibility requirements for video. The thinking is the same now: the feature should be in place so no one has to wait or build their own solution in the meantime.
Further ahead, the goal is to detect a file's origin automatically, based on the industry standards now taking shape. If a file was created by an AI service, it should be labelled automatically in Mediaflow.
Want to see it in action? Book a demo or get in touch with your contact at Mediaflow.
We spoke with product manager John Robertsson about the features that have already launched and those still to come during the rest of 2026. After a run of releases early in the year, he promises the pace keeps climbing.
A common problem for many people working in Photoshop is that the creative work happens in one tool while the source files live somewhere else entirely. Mediaflow's Photoshop integration lets you bring that whole flow into a single step.
Anyone who works with images and other media files runs into the same issue: when a workflow leans too heavily on manual habits, getting started as a new user can be tough. How do you quickly find your way through menus and folder structures that may not follow the logic that feels natural to you?