Q: How are publishers using AI in image workflows?
A: Working with various kinds of media houses, we find that publishers are very thoughtful about how to implement AI. It’s crucial to note that image workflows consist of multiple kinds of actions, some related to the images themselves, but just as many relating to their metadata. In the case of the latter, there are multiple ways of streamlining such processes with AI, for instance by automatically recognising key elements in a photo, adding relevant keywords and adjusting data values.
However, when it comes to the actual images, we find that modern publishers are more interested in disproving AI involvement, than relying on it. In a time when deepfakes can be created by anyone, publishers want to ensure that they can prove the authenticity of their images. This entails proving the image’s origin and lifecycle, and it’s made possible by implementing glass-to-glass C2PA (Coalition for Content Provenance and Authenticity) support to the content workflows.
C2PA is an international metadata standard that preserves cryptographic signatures embedded in a file, ensuring that viewers can trace the full provenance of the given image.
How, and to what extent, C2PA is supported will vary from one system to another, with the key element being that the signatures are kept intact throughout the workflow. Still, the real value of C2PA support comes when one can, in addition to not breaking the chain, inspect and prove exactly what has been done to the file. It’s not uncommon that an image is cropped or colour-adjusted before publishing, and small edits like these should be done without giving an indication of image manipulation. That’s why DAM systems (where images are kept during their lifecycle) should also allow for detailed inspection of the extent to which the file was edited, with which tools, at which times.
Q: What is a particularly illustrative use-case?
A: At Fotoware, the investment in C2PA support began with a request from a major UK picture agency, who were working on an initiative to prove the authenticity of an image. The ideal workflow should not only prove whether AI was involved, but provenance in general.
From capture to publication, the agency needed a solution that adds and maintains cryptographic signatures to each and every file, so that their provenance can be traced back to the original sources. While the rise of generative AI was the catalyser of the initiative, the ability to prove image provenance is also impactful outside of the AI-driven debate, as it serves a key role in rebuilding trust in media.
Early in the process, it became clear that every part of the workflow needed to follow the same standard, from the Canon cameras used by their team of photographers, to Fotoware, the digital asset management (DAM) solution. Not only would the DAM be crucial to preserve the cryptographic signatures but also to facilitate easy documentation of how AI was used.
As mentioned earlier, AI can play many central parts in an image workflow outside of image manipulation, and this is only expected to increase. For example, some publishers would want to implement third-party AI tools to read and / or populate metadata. Therefore, a DAM should be able to trace exactly how AI has been involved, and make this information easy to inspect. Additionally, such systems should ensure quick filtering, so that editors can easily filter based on the provenance records’ validity and / or AI involvement.
Q: What are your three top tips?
- Everyone working with content should be transparent in their use of AI. This is a legal requirement in the EU, but also an ethical and strategic choice long term. Trust is key to any publisher, and once it’s lost, it’s difficult to gain back. And while AI can help in many ways, it should not be implemented at the expense of content trust. Transparency, both internally and externally, is a key step to ensuring that this doesn’t happen.
- Publishers should keep the entire ecosystem in mind when they implement AI into their workflows. We know that for media houses, choosing best-in-breed software can offer many advantages compared to single-solution setups. Therefore, we recommend a structured approach that ensures the different systems adhere to the same standards, making both content and data available across the ecosystem.
- It’s important to have a broad view of the concept of automation: covering both AI and customised actions. It’s when one combines AI with tailored workflows that real efficiency is achieved. For example, publishers can use AI to recognise certain elements in a photo, then build an automation on top of this that routes files to the correct destinations based on what’s been recognised. As long as image provenance is kept intact throughout such workflows, the AI involvement would also be documented, and its role in the process would be clear to all.
About us
Fotoware is a Norwegian software company with 30+ years’ experience building digital asset management (DAM) platforms. Their first product, Fotostation, was released in 1994, serving as a solution for photojournalists handling large volumes of images. Today, Fotoware’s solutions cover a wide range of use cases for organisations worldwide. Contact Amalie Widerberg, product marketing manager.
Email: amwi@fotoware.com
Tel: +47 479 49 471
Web: www.fotoware.com
Janniche and the other contributors to our AI Special will be taking part in an ‘AI Special – Q&A’ webinar on 24th November. Click here for more information and to register.
This article was included in the AI Special, published by InPublishing in October 2026. Click here to see the other articles in this special feature.
