I’ve been having a lot of discussions lately with my father, a scientist and philosopher in his own right (lately, more philosopher than researcher, actually), about artificial intelligence. At 94, he still asks the right questions.
We already know the answer: AI invents nothing. It feeds on the knowledge we give it. You can use it for research, but it becomes truly useful once you already have an idea and want to develop it, structure it, push it further.
That’s exactly the spirit behind what we do at DIGITUM Consulting.
AI as a tool, not a replacement
We don’t ask AI to exercise artistic judgment or decide what deserves to be preserved. We ask it to apply, at scale and consistently, rules and standards we already know well: metadata standards, archival practices, digital authentication standards. It’s accumulated human knowledge that we’re putting to work to build tools that save precious time for photographers and collections.
We’ve talked about this at length in our previous articles, and a few ideas keep coming back:
Metadata is the memory of your images. Without it, a photo loses its context, its author, its provenance — much like the “Author” field, which remains the first line of protection for an image online. We’ve also pointed out that ten well-chosen keywords are worth more than a hundred poorly thought-out ones: quality of information beats quantity.
Most photo collections are at risk, often without their owners realizing it — poor backup practices, no structure, relying on a single external drive (the myth we debunked: the 3-2-1 rule exists for a reason).
AI remains an essential ally to the archivist, not something that replaces them. AI can speed up data entry and structuring, but it’s human expertise that ensures it all makes sense.
Digital trust — through Content Credentials, or complementary layers like Liccium — is becoming essential to protect the authenticity of a work in the age of AI-generated imagery. And that raises a question we’re increasingly asking creators: opt-in or opt-out when it comes to AI models training on your visual heritage?
And those who push back, sometimes rightly so
I also understand those who resist AI. Many see it as a threat, and the reasons are far from unfounded. There are real ethical issues: who has the right to use a work to train a model, without consent or compensation? There’s the environmental footprint of data centers, which is far from trivial. And there’s that broader fear that the technology will eventually outpace us entirely, even leading to our own downfall.
These concerns deserve to be taken seriously, not brushed aside. That’s exactly why, at DIGITUM, we insist on an approach where humans stay in control: we use AI to apply rules that we ourselves have established, not to hand it judgment or creative decisions. There’s a fundamental difference between being subjected to a technology and using it knowingly.
The idea stays human, the tool accelerates
That’s essentially what my conversation with my father comes down to: AI has no intuition, no lived memory, no aesthetic judgment. But give it a clear idea and a well-defined framework, and it becomes a formidable accelerator for developing it.
At DIGITUM, that’s the philosophy guiding the development of our tools: take established archival and photographic knowledge, and use it to build solutions that help creators and collectors better manage, protect, and bring their digital heritage to life.