News, views, & tips to do data science in byte-sized drabbles about my daily life as a data professional.

The deeper I go into developer and digital asset management work, the more I’ve gotten sucked into artificial intelligence. If you haven’t noticed, AI is and has been everywhere in the data field since computers came about. The rise in GenAI has created a lot of societal, political, and economic issues and I don’t use my newsletter to talk about those real, serious issues.

When I say AI, I mean the socio-technical issues with it like how AI is already in every single imaging device—from scanners, smartphone cameras, to digital cameras. There’s auto-white balancing, reading file formats, creating file formats, brightening/lightening, and more. This issue talks about AI in that sense and asks you to sit with the discomfort that it’s not just AI everywhere, but rather larger monopoly tech stacks like cloud-based infrastructures that pose similar socio-technical issues.

Does your company use a cloud-based database or server? It probably does.

Last year we saw a global Cloudflare outage that brought up some interesting DAM/CMS/database questions that I want to talk about it. Earlier in October 2025 we saw a major Amazon Web Services (AWS) outage (a lot of DAM/CMS' use AWS for deep storage). And now Cloudflare, which is a network security suite of tools for internet service providers, went down. Like AWS, Cloudflare is part of the underlying tech to the data systems we use for asset/content management.

Different DAMS/CMS' say they're different from other vendors by arguing they give different features/systems or targeting metadata standards for a specific sector. But this an illusion of choice. Most products will rely on the exact same underlying cloud technologies like AWS, Cloudflare, or SSL. The vendor we choose is really giving us front-end features (e.g. drag-and-drop file upload, list making) that may or not may not hold up when we export our data. The underlying tech to a system is more important during data migrations.

Many places use hosted options and therefore these technologies for reasons like lack of budget/funding or they lack in-house staff. The illusion of choice is important because there's continued interest in cloud-based systems, which makes us all susceptible/over-reliant on tech that DAMS/CMS vendor has no control over.

I attended the keynote presentation on creation, access, and re-use of digitized cultural heritage assets by the International Council on Archives back in December. A big point of the keynote was that a digitized image is a representation or interpretation of what the original material looked like shaped by certain decisions like file formats or resolution.

What stood out to me was when they said that AI is always used now to capture images. Smartphones and digital cameras all have automated AI processing for imaging (e.g. filter skin texture of people, make images more "clear", automatic white balance). In archives, we might "digitize" a research request for someone remote by taking a photo of it with our phone or tablet. Or we receive digital files from photographers for community/university/corporate events.

Our peers and researchers don't have that context to how digitization is a sum of technical decisions. They just see a photo. And most dialogue about digitization from people outside DAM/archives labor is about "why don't you digitize everything" or they conflate digitization with digital preservation.

How do you think about digitization of images? Do you document your technical decisions? How do you see automated AI workflows happening in your labor?