WHY PHOTOGRAPHY IS SO MUCH MORE IMPORTANT NOW IN THE ERA OF AI-GENERATED IMAGES.
Creativity and Craftsmanship are still kings!
I am going to talk about two AI-related big issues deeply embedded in the creative workflows in advertising and marketing agencies: CREATIVITY AND PROMPTING.
Issue Number 1: CREATIVITY
From” Andrea Laporta from 14 Agency, London: With AI in the mix – what still makes a photographer truly indispensable?”
1- “Even though AI is great, it’s still just a tool. It will never replace the input and vision that a photographer brings to a project. Image creation requires a level of sensitivity that only humans can understand and translate into authentic emotions for the viewer.”
Andrea’s opinion above is a certain realization that happens when we are basically forced to understand and assume that, as of today, everything that is created by any AI platform comes from images previously created, image portions and fragments of different exisiting images (both photographic and paintings) blended and reshuffled through several iterations, an attempt to prepare a different kind of salad made out of the same ingredients, and, -so far- this proves the inability of those systems to actually “create” something completely new, even less “translate into authentic emotions for the viewer.”
Whatever vision, craftsmanship, and ability to combine different techniques, from analogue to digital, to generative, all combined under each creator’s style, -basically the natural working realm of photographers- is still far from being disposable, removable or replaceable. Anyone claiming the contrary is just another conformist of the “Good Enough” gang.
Knowing so many may will question this argument under their own theories, which I am happy to hear about, etc, let’s jump to Issue number 2
2- Issue Number 2: POOR PROMPTING:
Now, putting aside this creativity argument for a moment, we have to enter the territory where many marketing and advertising agencies are obviously exploring the possibility of creating full visual campaigns from scratch, and thus saving a lot of production money, fees for talent and many other production values, plus post-prodcution expenses and time, and that might make many to think this is going to be the future of many projects.
The access to so many different AI platforms that can generate images has grown in a very fast and unexpected way. At the same time, it has brought a lot of unexpected problems because of the issues related to poor quality results linked to the massive lack of knowledge about photography and lighting of the “creators”, leading to inconsistencies that are hard to solve via iterations.
The main problem is the poor prompting, and the root of poor prompting comes from from not knowing the correct descriptions for all the related Elements of Visual Design ( form, shape, tone, texture, lines, patterns, color) all linked in one way or another to the basic compositional styles( Central Symmetry, Rule of Thirds, Fibonacci Rule, Dynamic Layering) for taking pictures, and even worse yet: the poor description of the light sources, Hot lights, strobes, LED’s. HMI’s, Kinoflos ( yeah, many are coming back to those!), the color temperature of the light sources, the lighting scenarios involving the manipulation of Shadow and Specular Edge Transfers, plus the way this lighting setups work in tandem with the diffferent elements of visual design described above, all these sets of parameters to optimize the impact of the client’s product in the final images, be it a still life or involving people or animals, or locations or all of the above.
No matter how cool anyone can “borrow” these concepts to generate “something,” the lack of actual experience in the matter will make a huge difference. There are already several examples to discuss this chaos, and one, with no names and no brands, will exemplify the situation:
A photographer is called to shoot 12 products in a classic still life scenario, clean background, making the product “float” in each image, all of it just following a clear brief. Art Director and Creative Director accompany the shooter, define colors for background, lighting, composition, plus all values described above. When the team is happy, all images get shot and delivered, BUT, the client shows up days later, saying they have 6 more products to shoot!
The ”creatives” at the Agency, decide not to call the photographer, save the time of production and financial expenses, as they will just do it all in AI.
Long story short, the agency wasted a lot of time and money to achieve a “good enough” result, which the client did not approve. The inconsistencies were notable and were not sorted out properly. Days later, they headed back to the studio to shoot the new products.
I am not saying that this type of workflow with good results is not achievable. It actually is!!
The problem is that the prompting process, the way to properly write in a way so the AI system clearly works according to the full control of all the parameters you want involved, not letting the AI “improvise” due to lack of concise information, brings out the reality of not knowing how to properly prompt when describing images in it’s obvious complexity, no matter how simple they look.
HERE IS A GREAT CHALLENGE I HAVE DEVISED TO PROVE THESE POINTS AND FOR EVERYONE TO GIVE IT A TRY:
3- Grab an image you like, or even an image you dislike, it may be advertising-related or just any simple snapshot of familiar places, a picture of your favorite artist, just any image and sit down to write a text describing in full detail everything you can tell about the image, from the subject, to the composition, to the lighs, the shadows, the lighting, the textures, the tones, shapes, etc, etc.
Once you have done your “homework,” proceed to drop this text, this prompt in any or even all of the familiar AI systems available. Most of them have a free version online, so you can drop this prompt into Nano Banano, Claude Online, Chat GPT, Grok, the Chinese one and even Perplexity.
Now, if possible, print the results, and sit down and “enjoy” how close or how far the images created by the AI systems are from the original picture, based on your description
This is the perfect, simple way for you to establish how cool your prompting skills are, how knowledgeable in terms of photography and lighting you are, and this will inform you of the necessity of having (or not) a photographer handy to assist you in the prompting stage.
Now, THAT is how relevant photography is in these days of AI image generation.
You can, of course, extend this to movies and TV commercials, and if you are shooting an AI-generated commercial video, then a good cinematographer may be your new BFF!
Cheers!!







