The building itself looked like it had been making things for over a hundred years, probably because it had. Multi-story brick, freight elevators, the kind of Chicago industrial space with the heavy wood beams and wood floor. Nobart occupied the entire building, a catalog production house at the absolute top of its craft in 1987, running 8x10 view cameras under tungsten hot lights, with its own photo lab and production department. It had a workflow refined to the point where every decision had a reason and every reason had a history. It had to be if you had a client list that included Sears, Marshall Field’s and J.C. Penney.
I was there as a visitor, not yet a full-time working photographer, and what I remember most is the gritty feel of this working studio. The cameras were enormous. The lights threw heat you could feel across the room. The images were technically flawless in a way that required genuine expertise to achieve, and everyone in that space knew their craft.
Tucked in a corner of the production department, right alongside the keyliners, strippers, and typesetters, almost as an afterthought, was a Sony Mavica still video camera being used to shoot small fabric swatch insets for the very same catalogs being worked on upstairs. What a curiosity, a workaround for a specific low-stakes problem. The dominant workflow carried on upstairs with complete confidence. Why worry?
The disruption wasn't announced. It was already in the building.
To understand why that Mavica mattered, it helps to understand what photographers had already lived through.
Every major disruption before digital followed the same pattern. Automatic exposure removed the need to meter by hand. Autofocus removed the need to focus by eye. TTL flash exposure removed the need to calculate exposures by hand. Each wave took something that required technical skill and made it invisible, and each time photographers adapted by moving toward the work that still required judgment - composition, lighting, the relationship with the subject, the decision about what the image was actually supposed to do.
The lesson the industry internalized, not always consciously, was that technology handles the how and the photographer owns the why. Every disruption that automated a process created space for the photographer’s vision and voice to move up in importance. Sometimes it was uncomfortable learning the new tech, but each time the professional found their footing.
This is the framework most working photographers carry into the AI conversation. It's a reasonable frame. It's also incomplete.
The catalog disruption of the late 1990s looked like a technology story. It wasn't, or at least that wasn't the whole of it.
Digital cameras were improving, yes. But the work didn't leave independent photographers because digital was better. It left because prepress and printing companies needed to support their core business in the fast-changing industry – keep the presses running and the staff paid. Bundling photography into their services solved that problem. They weren't better photographers. In many cases they were significantly worse. But they had a different reason to want the work, and they had the client relationships and the pricing leverage to make it happen. If needed, they could give the photography away to keep the valuable print job.
The technology was the trigger. The real force was an adjacent industry player with a different set of economics and a specific incentive to absorb the photographer’s work into theirs. To them, setting up a digital photography studio from scratch and hiring someone to run it was a rounding error when it came to their bottom line.
This is a pattern worth naming because it repeats. The most dangerous disruptions rarely announce themselves as disruptions. They arrive looking like a service offering from someone you already do business with, priced in a way that's hard to argue with, solving a problem the client didn't know they had. By the time the mechanism is visible, the work has already moved.
It happened with stock photography platforms absorbing assignment work. It's happening now with AI image generation being bundled into the design tools clients already pay for — the trigger changes. The pattern doesn't.
But even when the mechanism was visible, timing remained its own trap. In 1992, I was assisting on a location shoot in Indianapolis. We were working on a medical supply catalog, a telephone-book-like catalog that we had two days to shoot a slew of product groupings. I was traveling with a photographer I was working with regularly, and the designer who had come along, who was, to put it charitably, evangelical about the future.
The camera was a Kodak DCS 200. A groundbreaking 1.5 megapixel wonder with a 2.63x crop factor and a mechanical hard drive built into the camera body that you could hear spin up as you turned the camera on and it would click away when it wrote each frame. The camera was tethered to a Mac Quadra via SCSI cable, the whole rig was fascinating to look at and the designer was absolutely certain it was the future of photography. He said so repeatedly. “Film is dead!” He said it the first day and the second day and at various points in between, with the particular conviction of someone who has seen something true and cannot understand why everyone around them isn't responding accordingly.
He wasn't wrong about the direction. He was wrong about the timeline, and in the gap between those two things lay a great deal of damage for photographers who listened too carefully to people like him.
In that bit of technical wonder, a disaster waited quietly. In the designer’s excitement to work with the future, he would play with every image we took, which was stored on the camera’s hard drive, draining the battery faster than the power system could keep up. It didn’t take long for the camera to die and the computer and camera would lock up. For those who didn't work through that era, a SCSI or “scuzzy” cable was a data connection with very specific ideas about how it should be treated, none of which included being disconnected from the camera in frustration. Every time the system would lock up the designer would unhook the computer from the camera. After the fifth or sixth time of this chain reaction, his Mac decided it had had enough and revolted. The system OS became corrupted. He spent the better part of a day and a half on the phone with Apple tech support in Cupertino trying to resurrect the system, while the photographer and I sat twiddling our thumbs, watching the future malfunction in real time.
The most vocal prophet of film's death was undone by a cable he shouldn't have pulled.
Direction and timeline are different things. Photographers who panicked at the evangelical certainty of people like that designer made expensive decisions years before they needed to. Photographers who dismissed the direction entirely made equally expensive decisions in the other direction. The ones who navigated it best held both truths at once — yes, this is where it's going, and no, it is not going there yet.
That discipline — seeing clearly without reacting to the hype — is exactly what the current moment requires.
When the pendulum swings, it swings hard.
A photographer I know lost a significant tool catalog to exactly the kind of bundled service described above. The printer absorbed the work, staffed it with people who could operate the equipment, and the client accepted it because the price was right and the product looked adequate. Sufficient. Good enough.
Two years later, the client called back. The vendor had been handling the single and small group product shots reasonably well. But the catalog included sets — tool groupings of five hundred pieces and more, arranged to show system and relationship and hierarchy, the kind of shot where every placement decision carries visual logic — where a 500-piece set has to read as a system, and a 1000-piece set has to make a thousand equally important components feel like a coherent whole.
The new people couldn't do it. Not because they lacked equipment, but because they lacked the judgment that only comes from years of understanding what a complex arrangement is actually supposed to communicate.
The market had confused production capacity with expertise. The client paid the price and knew it.
This pattern — overcorrection followed by recalibration — has repeated across every disruption I've watched move through this industry. It showed up again around 2012 when iPhone photography became capable and clients began wondering aloud whether they needed photographers for certain work at all. Some didn't, for certain work. But the recalibration came, and when it did it came in the form of clients saying plainly that they didn't want that iPhone look anymore — that sufficiency had quietly become a liability rather than an economy.
“Good enough” is a moving target. The market finds it, uses it, and then some notice what they lost.
Here is where we need to take a harder look.
The frame established earlier, that technology handles the how and photographers own the why, has held through every disruption described above because every previous disruption automated execution. Metering, focusing, flash calculation, even the digital darkroom. Technical tasks, however skilled, could eventually be systematized. The why remained with the photographer because the tools had no access to it.
AI is different. Not incrementally different — categorically different. For the first time the disruption reaches in to take the photographer’s judgment itself. AI can now generate images from text descriptions, make complex compositional decisions, simulate lighting, re-light existing photographs, and produce work that is not merely sufficient but, in certain contexts, seemingly accomplished. The how and the why are no longer cleanly separable in the way photographers have relied on for forty years.
This is worth saying plainly rather than explaining away. The photographers who navigate this best will not be the ones who reach for the reassurance of old patterns, insisting this is just another Kodak DCS moment, another pendulum that will swing back. Some of it will swing back. Not all of it.
But — and this is the distinction that matters — encroaching on judgment is not the same as replacing all judgment. AI can simulate the decisions a photographer makes. It cannot be present. It cannot build trust in a way that changes what a subject is willing to show. It cannot carry the intention that makes a portrait something more than a technically accomplished likeness. It cannot mean anything to the person sitting in front of the camera. In short, AI cannot hug your clients.
The lanes where human presence, relationship, and intention are not incidental to the product but are the product. Those lanes are not under the same threat. The question is whether you are working in one of them, and whether you can turn it into a clear client benefit.
The response to sufficiency was never better equipment. It was a deeper intention.
I felt the disruption of the iPhone differently from the disruption of the catalog work. The work that left wasn't complex tool sets that required expertise to arrange. It was the casual family snapshot, the quick headshot, the image whose only job was to document a moment. Sufficient was genuinely sufficient for that, and pretending otherwise was a losing argument.
But something else was becoming visible at the same time. In a culture that produced over 2 trillion images in 2025, which is just shy of the total number of images taken in the 150 year history of film, the photograph that means something is not the one that was easiest to make. It's the one that was made with intention, for a specific person, at a specific moment, that someone recognized was worth pausing for. Not technical capital, emotional capital.
This is one answer. The principle underlying it applies more broadly: when sufficiency is available everywhere, presence and intention become differentiators only for photographers who have deliberately built them into their practice.
The portrait session I offer doesn't begin when the camera comes out. It begins with a conversation — not only about clothing and location, but about what prompted the call. What are you noticing? What feels different? What do you hope to remember about this particular moment? During the session, the attention is on how people move with each other, where ease shows up, and what genuine connection looks like when it isn't being performed for a lens. The portraits matter. But they're not the only thing being made — the session itself is the work, the relationship its medium, the print its residue.
My paper negative portrait project extends this into territory that makes even the process itself a shared discovery. A large, unusual camera. A slow, uncertain medium. A process that disrupts what the subject expects to happen, and in that disruption — in the errors and the serendipity of a difficult analog process — reveals something neither photographer nor sitter could have planned for. The uncertainty is not a limitation. It's the point.
This is not a nostalgic practice. It's a deliberate answer to a specific market condition: when sufficiency is everywhere, presence and intention become the product.
Every disruption described here arrived with prophets. Some were right about the direction and wrong about the timeline. Some were right about the technology and blind to the real mechanism. Some were loud enough and convincing enough that photographers made expensive decisions based on the certainty of people who turned out to be yanking cables they shouldn't have pulled.
The pattern that emerges across forty years is not that disruption doesn't matter. It does. It's that the photographers who navigate it best are not the ones who react fastest. They're the ones who read it most clearly — who separate direction from timeline, who identify which adjacent player has an incentive to commoditize their work, who understand which parts of what they do will remain despite the disruption.
AI is the most complex disruption this industry has faced because it reaches further into the work than anything before it. That's worth taking seriously. The adjacent player who bundles your work into theirs hasn't gone away. They now have a new tool. It's also worth noting that the lanes where human presence, relationship, and intention are irreducible — where the experience of being seen is itself the product — are not the lanes most at risk. They are the lanes most worth building.
The question the current moment asks of every working photographer is not whether AI will change the industry. It will. The question is whether you are making deliberate choices about where you stand, what you offer, and why it matters to the specific clients you serve — or whether you are waiting to see what happens and hoping the market leaves your corner of it alone.
Pick your lane deliberately. Before the market picks it for you.
Editor’s Note: Originally, I planned on presenting this post without images since most of what I am discussing relates to cameras and technology I don’t have pictures of, and finding those images online might create copyright issues for our association. I was encouraged to post images with the post… The images I have posted are some of the earliest I have access to. They are from a Kodak DSC 560c, a 6MP camera built on a Canon EOS-1n body with a 1.32x crop, and used a PCMCIA card for image storage. The selling price when introduced in 1998 was $28,500. The camera recorded the raw images in a proprietary TIFF file that was an early raw format that required Kodak software to convert. I reprocessed these images, which were taken in early 2000, from the original TIFF files, converted to DNG files. Converting 25+ year old raw files from this camera was a trip, not nearly the same headroom that is available with today’s files.
