the role(s) of AI in Photography
Like it or not, Artificial Intelligence has become a seamless part of our daily lives. The systems we rely on are constantly evolving, making our experiences smoother and more intuitive. Yet, when it comes to photography - and the arts in general - AI remains a hotly debated topic, often dividing opinions.
A digital image created during early experiments with AI.
In this episode, I’d like to share my personal thoughts on the current state of AI in photography and how I’ve come to embrace it in my own work. Of course, these are just my personal views, and I completely understand that others may see things differently.
AI Content Generation
To dive deeper into this discussion, it helps to have a basic understanding of how AI works - at least from my perspective.
At its core, AI generates content by analyzing vast datasets of text, images, and other media. When creating new work, it pulls from a wide range of external sources to produce results.
For example, when prompted to generate an image, AI sifts through these datasets, identifying relevant patterns and details to construct a visually representative output.
This process differs from making edits, such as removing distractions from an existing image. In that case, AI relies on the image itself, using its existing data to create a seamless adjustment.
A hybrid approach comes into play when adding new objects. Here, AI blends external object data with the original image, ensuring the addition feels natural and cohesive.
The Contrasting Roles of AI in Photography
In photography, I see AI playing two distinct roles:
as a creation mechanism – Generating output based entirely on a prompt typed into a computer.
as a support tool – Refining and enhancing an existing piece of work.
AI as a Creation Mechanism
This will be quick. I don’t believe that an image generated from a computer-based prompt qualifies as photography in any way, shape, or form. To me, a photograph must come from a device that captures what is in front of it - essentially, what a camera records.
That said, I have no issue with AI-generated images; I just see them as belonging to (and deserving of) a separate artistic category, distinct from photography.
Beyond image generation, AI can also create text narratives around photographic work using tools like ChatGPT - a topic I’ll touch on later.
AI as a Support Tool
In this role, AI offers many valuable features that photographers can leverage.
Image Enhancement Support
The use of AI in image editing is a hot topic, but context is key. My stance depends on the intent behind the photograph and how AI is applied. When it comes to documentary-style images, any manipulation should be done carefully to ensure the image stays true to the scene as it was observed.
However, for image enhancement, I fully support the benefits AI brings. In Adobe products like Photoshop and Lightroom, tools such as Generative Fill can produce remarkable results, especially when removing blemishes or distracting elements from an otherwise unusable image. To me, this is simply the next step in the evolution of past tools like Content-Aware Fill and the Cloning Tool, which served the same purpose.
For purists who argue that AI undermines photography’s integrity, I’d counter with the fact that photography has always involved a degree of manipulation. Photographers of past generations used darkroom techniques like dodging and burning to refine their images and carefully selected film stocks to achieve very specific results. Even today, choices in lens, aperture, and shutter speed allow photographers to shape their images in ways that enhance artistic appeal.
The extent of AI manipulation in an image remains a personal choice. Is sky replacement ethically acceptable? What about adding objects into a scene with AI? I haven’t explored these possibilities extensively myself, but if the goal is to create an artistic interpretation of a scene, I believe AI has its place.
In this more extreme example, a distracting figure at the edge of the left image was removed using AI, saving it from deletion
Analysis Support
In an earlier blog post, I talked about how ChatGPT can now provide an unbiased critique of an uploaded image. This can be really useful for photographers, as it removes the subjectivity that comes with a human assessment.
In a follow-up post, I shared the results of tests using ChatGPT to evaluate websites. Useful in its perspective elsewhere, this can be invaluable is its assessment of a personal website. In my case, the tool provided a surprisingly accurate overview of my own site, and better than I could have worded myself.
Although not a core strength of AI, I believe this objective analysis can be helpful and put to other uses.
Text Support
Writing isn’t exactly my strong suit, so I tend to lean on ChatGPT quite a bit for support. That doesn’t mean I use it to create entire documents—though occasionally, I’ll ask for ideas I might not have thought of. Instead, I run my drafts through the tool to check grammar and get suggestions that could improve readability.
The key thing to note here is that the draft is always written in my own voice first. When I use ChatGPT, I’m simply looking for ways to refine it. I then pick and choose which suggestions enhance the text while keeping my original intent intact.
And yes, this post was fine-tuned with a little help from my ChatGPT friend! 😊
Final word
AI-generated works are inherently derivative, built from a pool of existing creations rather than the unique vision of an individual. Due to a dilution of styles resulting images lack any distinct characteristics or personality of an individual artist. As such, AI-generated content can often feel soulless and impersonal.
That said, I find AI to be a valuable support tool when used with a light touch. It’s incredibly useful for repairing blemishes in images, and ChatGPT helps refine my grammar in written passages. The key is to ignore suggestions that stray too far from my own voice.
Make no mistake, AI-generated images can be astonishing in their own right. The controversy surrounding this art form doesn’t come from the tool itself but from how it’s presented. Too often, AI-generated works are shared by artists who knowingly allow the public to believe they were captured by a camera. The stigma around AI will persist until clear standards require AI images to be labeled as such - something that’s already possible, given the absence of EXIF data.
Personally, I see real value in using AI tools to refine images (particularly abstract ones) and enhance text that supports my work. I’m always mindful to keep AI under my control and ensure that any changes align with my creative vision.
I can see a future where AI allows me to upload a representative body of my work and from that dataset generate images that more closely reflect my personal style. But that day is not here yet - and when it does arrive, it will present even greater ethical challenges for photographers.
With my thoughts on AI out in the open, I’d love to hear your take on its role in photography! Do you use it to create or edit images? Have you tried using ChatGPT to craft supporting narratives or analyze your work?
Drop your thoughts in the comments below, I’d love to hear your perspective! 😊