During these years dedicated leading photo tours to Cuba, teaching and photography in Cuba, I have had many conversations with my travelers about AI.
Here I show you some of my Cuban photographs generated with AI and thoughts, analysis and reflections around this technique of image generation.
My first thought about AI
My first thought when I started working with artificial intelligence in my cuban essays was "Hey man, these are not photos, they are images, just pixels".
I am a classic documentary photographer, I find the magic of photography in the act of thinking, searching and finally pressing the button and getting what I now call "an authentic photograph".
His name is Alvaro, he lived and died.
Until 2023 all my cuban photo-projects have been done with cameras, travelling hundreds of kilometers with my legs, my camera, my cap and my water bottle.
I know the names of many of the people I have photographed after chatting with them (a great advantage over AI), this human side of photography is what made me fall in love from the first moment I had a camera in my hands and I believe I will continue to be hooked to this drug of "knowing that the people I have photographed are real, they have life and I am a spectator of this great play that is life".
AI for projects, not for single images
At the moment it does not occur to me to use AI to obtain single photographs of Cuba, I mean I find it very attractive to generate projects such as recreating Havana in the 50s.
Maybe it's because as a documentary photographer I'm more interested in history than in the beauty of a single photo.
There are many essays that I am interested in generating, stories where I could not be present such as: slavery in Cuba, the life of Cubans in communist Russia and a long etcetera of projects that I don't know if I will create since my true love is the photography made in the street with my camera.
More editors than photographers
The results of using AI for image generation are always surprising, in my research on Havana in the 50's there were images that surprised me as the software interpretation is always a surprise.
Eventually I realized that I was trying to improve the text to get a better image, but in the end it was a matter of faith, waiting for images worthy of entering my 1950's Havana project and finally placing them in the right order to give visual and narrative coherence to this digital essay on the old Havana.
In short, a job more of an editor than a photographer
Are you a photographer but never use your camera?
This is the big question. Until 2022 we would never have imagined a person who would not get up from his chair and that the same "photographer" would be called; perhaps now we will have to redefine the concept.
My opinion is that perhaps new terms need to be coined, but for me a documentary, or travel, or street photographer needs a camera, a brain behind the camera, and good shoes.
Perhaps we have to distinguish between a photographer and a digital artist, between a classic photographer and an AI image generator, between a photographer who walks and a photographer who doesn't walk.
The passage of time will help us create new terms to define these new jobs and concepts.
IA's pessimism for the Cuba of the future
Generating images with AI related to Cuba is surprising, sometimes thoughts pass through my mind about how these programs "imagine" Cuba in the future and the results are very curious since they are usually pessimistic, the destruction continues, the broken elements, the old cars.
If we do the test with other countries the results tend to be futuristic in a good sense , but not in Cuba.
It is sad and disturbing how these softwares do not "trust" the progress of this country.
There are many and varied doubts and thoughts that come to my brain since I discovered AI applied to photography.
Here are some of them.
* Is it correct to use AI in documentary projects?
I think we could not call it 100% documentary since it is a software that has "documented" it, maybe we could call it "documentary-artistic-invented", who knows !
* When we are missing a photograph in a project and it is impossible to take it.
Can we generate it with AI in order to have a more complete project?, honestly I have to think about this, I do not see it as a pure option, but it is very attractive the possibility of finishing those projects that are half finished because some image is missing.
* This is the work of a photographer or an editor? after conversations with other colleagues I am inclined to think that when using AI to generate images what we are doing is waiting for the "appearance" of images that we like and once we have them we start to place them where we want, so I think that in many occasions we are more editors than photographers.
We are just waiting for the correct image, "probably we are more writers and waiters than photographers"
Why can't I do a project on slavery in Cuba?
One of the topics I have read the most about has been slavery in Cuba, the years of the slave ships, the exploitation of sugar cane, etc.
It is a topic that I would like to work with AI but it is almost impossible because these software interpret that there are painful elements, or that can offend sensitivities, therefore I cannot start this project that interests me so much.
Perhaps in the future AI will allow projects to be carried out as hard but real as the exploitation of people for the benefit of a few.
Trying to rejuvenate elderly people I met
I'm trying to work with AI on my "old hats" project, they are older Cuban men wearing their old hats and caps.
My intention is to rejuvenate these people and see what the software's interpretation is of what that person was like 50 or 60 years ago. The results I am obtaining are satisfactory in terms of interpretation of the person's clothing and position, but it is almost impossible for the face to look similar.
I think it's still too early for AI to do a proper rejuvenation of a reference photo.
What is a Cuban photographer for the IA?
I have been taking photo tours in Cuba for many years and it continues to surprise me how my students, travelers from various countries, are surprised that in Cuba the population is not predominantly black.
This is because we all imagine Cubans watching movies, clichés, etc. but the reality is that 70% of the population is white.
Here I show an image generated by artificial intelligence in which I request that you create an image of a typical Cuban photographer.
Surprised?
Do you understand the IA of photographic composition?
Photographic composition has many rules, but the experienced eye of photographers makes us see it without thinking about these rules.
Perhaps a good composition is what I like most about a photograph, and here I have a question about how much the AI knows about photographic composition.
I have done the test and this is the result of what I consider to be a good composition in street photography in current Havana.
The best photographs with people generated with midjourney are usually with a person in the center, so it seems that this is what you understand as "good composition".
It is demonstrated then that Midjourney does not yet speak the language of photographers perfectly.
Past, present and future of Cuba (according AI)
For many years I have heard travelers who participate in my photo tours say that phrase "today's Cuba is stopped in time", it is one of the great attractions of Cuba, the curious thing is that the forecast for the future that the IA programs make is almost null.
I mean that they continue to see Cuba with the same topics, the same cars and the same cracks in the houses. A pessimistic version that I hope there is a way to change, both in Midjourney and in the future development of Cuba.
Beauty is nothing with AI
Creating a portrait of a beautiful person using AI will take 1 or 2 minutes, so I personally believe that there is not much value in this type of image (not photography).
I could make a portfolio of Cuban models in a morning but its value to me is null, it is not my photographic inclination.
I understand that studio photographers save a lot of money thanks to AI and can expand their portfolios with beautiful images of "models", but once again the same doubt arises.
Are they photographs or just pixels?
Advantages of AI in photography
* Undoubtedly one of the advantages of AI is to carry out visual projects where it is not possible to be, either because of the place or the time in which they happened.
Example: Havana in the 50s
* We can imagine the savings if we don't have to pay for models, lights, camera, travel, hotels, etc. Stock photographers have an important saving, besides, every day it is more and more clear that AI generated photos will be accepted in image banks.
* We don't need the model release signed by the people we photograph, because as we know "they are not real people, just pixels in the shape of a person".
Disadvantages of AI in photography
There are certainly some disadvantages related to AI in imaging.
Many illustrators have lost a lot of business since people can generate a lot of images themselves.
Regarding photography there are certainly disadvantages as well, here are some of them.
- Many models lose their work as their presence is not necessary.
- The generated images are closer to art than to photography itself.
- Photographers are tempted not to leave home and end up being digital artists generating images.
- Everything is confused, we will no longer know if a person is real or not, maybe for the final public it is not important, for me it is important because I practice photography because I consider it is a key to approach and understand the human being.
- A big disadvantage is that anyone can copy your image as they can duplicate your prompt or make an image based on yours so this takes away from the strength of your creation.
Imagined world that feeds on itself
With the current realization of these works done with AI we are generating a large amount of images that will serve to be the model for the next generation of images with AI.
In short, we are creating an imagined world that feeds on itself and that will become increasingly larger and certainly less real.
AI and Photoshop in my images of Cuba
As I tell my students in Cuba, I believe that there is no need for much retouching if a photo is good, with this I mean that I hardly use Photoshop in my documentary photos (I retouch them slightly in Camera Raw), faithful to this philosophy I am not retouching the images generated with Midjourney either.
Actually if they have some small flaw I leave it visible, it is the signature that it has been made with AI instead of being taken with a camera.
Cuban photographers using AI
At the moment I don't know of any Cuban documentary photographers who are working with AI.
It is true that these are classic photographers.
In the world of fashion, nude and beauty we do have some reference as Izuky who always bets to be aware of the latest trends and who is experimenting in this new world.