Will AI really be able to replace Photographers…?
There’s a lot of talk about AI at the moment and it’s hard to know what to make of it.
In many respects it’s already very much part of of our day to day lives, operating behind the scenes in aspects of our lives that we’re entirely unaware of. And like all new technology it can be both a positive and a negative force, depending on who’s behind it, how it’s applied and the way in which we frame our understanding of it. I don’t think there’s much doubt that in scientific fields, especially medical research and diagnostics AI can help improve and even save lives through the identification of health issues long before any of the best clinicians in the world could ever hope to, and with an impressively high success rate. On the other hand, our general lack of deeper understanding tends towards viewing AI through the Sci-Fi lens of Philip K Dick or William Gibson novels, which in many respects are perhaps not that far from reality, but do tend towards a more dystopian interpretation.
Arguably, the Djinni is out of the bottle and whether we like it or not, we need to master AI before it masters us. The problem however is that the technology is developing at a pace too fast for our understanding. According to the American sociobiologist Edward O Wilson, “The real problem of humanity is that we have Paleolithic emotions, medieval institutions, and god-like technology.” I can’t think of an instance where this quote is more relevant than in relation to Artificial Intelligence.
There is a very real concern nowadays about the veracity of information. Deep fake videos, AI generated images, fake news and conspiracy theories can combine to leave people unable to tell truth from fiction. The old adage of “seeing is believing” no longer stands.
Unsurprisingly, many people have a genuine and largely justified fear that their jobs will be replaced by AI. Like every working photographer, I already use software which contains powerful built-in AI tools that assist in my workflow. I use a software called Capture One Pro which does everything a photographer needs in post production and has increasingly introduced more AI tools with recent updates, providing huge benefits in workflow. Similarly, Photoshop and Lightroom from the Adobe suite have incorporated equally powerful tools in their software which work in very similar ways. In my own workflow, these AI tools are primarily used to create masks (a technique for making targeted adjustments) applying consistent lighting and colour corrections across multiple images, removing blemishes and smoothing skin tones and so on, all processes which a couple of years ago would have been difficult and time consuming but can now be done in a fraction of the time.
As a photographer I thought I should dip my toe in to the AI waters of actual image creation to try and find out what it’s all about and whether or not I should be afraid. But where to start? The obvious starting point (for me at least) was the corporate headshot. This is an area where AI could quite easily do away with the need for a skilled human photographer. There is an understandable temptation for companies to try and save time and money and let AI generate corporate headshots. In fact I was recently sent a link to a Linkedin post that provided a step by step guide to writing prompts to generate a corporate headshot using ChatGPT with the promise of saving companies money on hiring a professional photographer.
So I gave it a go.
Without much preparation, I took a few selfies on my phone and uploaded them to the free online version of ChatGPT 4o.
Selfies taken on my phone
And of course I had to write a prompt, so I entered the following:
“Generate a corporate headshot of this person, wearing a dark blue tailor made suit, white shirt with no tie, against an out of focus background of a generic city skyline.”
I waited while the AI did its job (which took longer than expected, and probably contributed to the massive detrimental environmental impact that AI has in the process, for which I do feel slightly guilty but that’s a different issue altogether…) and eventually this is what appeared:
AI self portrait
I honestly had to control my urge to laugh out loud when I saw the result!
In fairness, the AI did a pretty good job in response to the very basic prompt I gave. But from a photographer’s perspective, it’s obvious that this is an AI image. I showed it to my wife and she could tell it was supposed to be me but agreed that it bears only a passing resemblance. My face looks rounder than in reality and it’s given me less of a receding hairline, and a perfect complexion. OK, so it’s not all bad! But more importantly, it lacks any of the qualities you’d expect from a genuine portrait, it’s not an authentic image and I can’t recognise any of the unquantifiable, intangible qualities I would expect to see in a genuine portrait. Although it’s done a good job based on the prompt for the background, the finished image looks very much like the portrait has been cut out and stuck on a different image, which of course it has, not to mention the lighting between the shoulders and the face is inconsistent, and the shirt and jacket don’t look real.
The problem is that many people may not be able to tell the difference between a genuine portrait and an AI generated image. Perhaps more worrying is the fact that some wouldn’t care and many would even opt for the AI image, preferring an artificially perfect representation over the true, imperfect reality of our authentic selves.
For me, that’s the key point; authenticity and in particular the value we as a society place on this concept. In a world with an ever-increasing amount of fake or AI generated information, I believe it’s more important than ever to represent a genuine version of ourselves to the world. Certainly in a business context I believe that authenticity carries a lot of weight. Trust is at the heart of all relationships, both personal and professional, so what message does it send to your customers or clients if you aren’t presenting your true self, but instead offering an artificial version in the belief that your clients will trust that what they’re seeing is the real thing?
I wondered what argument the AI itself would make of this issue of real versus fake imagery.
So I wrote another prompt:
“Write a 250 word blog post about why AI generated images will never be as good as images created by a photographer with a camera”
And here is the AI generated response:
Why AI-Generated Images Will Never Match the Power of Real Photography
“In an era where artificial intelligence is reshaping creative industries, AI-generated images have become a hot topic. While these tools can produce impressive visuals quickly and at low cost, they still fall short of what a skilled photographer can capture with a camera.
Photography is rooted in human experience. A photographer doesn’t just point and shoot—they interpret light, composition, timing, and emotion in real time. They work with the imperfections of the world: the weather, the subject’s mood, the play of shadows. These unpredictable elements give photographs a sense of authenticity and spontaneity that AI can’t replicate.
AI-generated images, by contrast, are simulations built from vast datasets. They may look polished, but they often lack emotional depth and real-world imperfection. There's no human intuition behind the lens—just algorithms trained on patterns. They can’t sense the moment a subject’s expression shifts or the way golden hour light hits a surface.
Moreover, photography tells stories tied to place and time. A real photo is a moment lived and documented. AI fabrications, while visually compelling, are disconnected from lived reality.
That’s not to say AI has no place—it can be a powerful tool in design or ideation. But when it comes to capturing the soul of a moment, the artistry, and the real-life presence of a subject, nothing compares to the human eye and the photographer’s lens. For that reason, AI will always be a tool—not a replacement.”
I think the above makes some strong points, especially within the limited scope of the prompt (250 words). But like the image itself, I think it’s obvious that it’s AI generated, it lacks the character of a human author. Where my view does align with the response is in its closing statement, “AI will always be a tool - not a replacement”. The challenge, not only for photographers is to learn how to master this powerful tool, to find and develop ways of thinking and working that will ensure that AI doesn’t replace humans.
But I also wonder how convincing an argument the AI could propose in its favour…
So I asked it:
“Write a 250 word blog post about why AI generated images are better than images created by a photographer with a camera”
And the response:
Why AI-Generated Images Are Outpacing Traditional Photography
“In today’s digital age, artificial intelligence is revolutionizing visual content creation—especially through AI-generated images. While traditional photography has long held its place as the standard for capturing reality, AI offers unique advantages that are reshaping how we think about imagery.
One of the biggest benefits of AI-generated images is their limitless creativity. With a few prompts, AI can create surreal landscapes, futuristic cities, or fictional characters—none of which exist in the real world. Photographers, limited by physical constraints, weather, location, and subject availability, can’t easily match this kind of imaginative output.
Speed and cost-efficiency are other major advantages. AI tools can generate high-quality visuals in seconds without the need for expensive equipment, travel, or post-production editing. For marketers, designers, and businesses working under tight deadlines or budgets, this is a game-changer.
AI also democratizes access to professional-grade visuals. You don’t need to be a trained photographer or hire one—you just need a creative idea and the right tool. This levels the playing field for creators and small businesses alike.
While traditional photography excels at capturing real moments, AI-generated imagery offers unparalleled flexibility, efficiency, and creative freedom. It's not about replacing photographers, but rather about expanding what's possible in visual storytelling. As the technology continues to improve, AI-generated images will only become more lifelike and accessible—further cementing their place in the future of digital content.”
Also in this instance I think the AI makes a good case, which is not surprising as it is analysing the vast amount of data available and generating an argument based on what it finds. I’m reassured by its insistence that “it’s not about replacing photographers”. AI acknowledges the difference between reality and a manufactured facsimile, and it recognises the value of “real moments”, which it acknowledges it cannot capture. I think this reinforces my point about trust and authenticity, and once again I’m drawn to the view that AI isn’t all bad and out to replace us, rather it is a tool that needs to be understood and applied with careful consideration.
So on the basis that any AI system is only as good as the input it receives, I decided to give it another go, to test its consistency and see if the results would differ at all. As my first experiment was done using a basic phone selfie I took another self portrait a day later, this time with a quality compact camera (Canon G1xiii) mounted on a tripod. I also wore a different pair of glasses to see if that would influence the process. Here are the images that I uploaded to ChatGPT:
Selfies taken with a compact camera
The Prompt:
“Generate a corporate headshot of this person, wearing a dark blue tailor made suit, white shirt with no tie, against an out of focus background of a busy office”
Again I waited longer than anticipated while ChatGPT did its thing and here is the second AI generated corporate portrait alongside the first for comparison:
First AI Generated Portrait
Second AI Generated Portrait
My reaction was pretty much the same as the first time; trying not to laugh! The AI clearly generated a very similar looking individual, with a vague likeness to me, but again, it altered the shape of my face and this time even changed the colour of my eyes! In both cases, the clothing and the glasses look unconvincing to me (especially the glasses) and the overall look is too polished, hyper-real in a way that leaves me feeling somewhat disconcerted and uneasy.
AI is of course here to stay and looking forward, there’s no doubt that it is a tool that must be accepted and indeed mastered. If that’s still possible. It is still in its infancy and in the long term, there may well be a time when photographers and many other professionals are replaced by AI. In the short term however, it’s less a case of AI taking people’s jobs than people who know how to use AI taking those jobs. And based on my first brief experiences with my self portrait experiment, I don’t feel the threat of replacement by machines.
Yet.