Did artificial intelligence technology take over photography? No, it hasn’t, and it probably never will. But here is how AI might have been helping you behind the scenes all along.
Artificial intelligence, or AI, is perhaps the biggest buzzword in the world of photography and tech for 2023, and it continues to be a widely discussed topic in 2024. This stems from the fact that the industry has had mixed reactions about the emergence of AI tools for photography and raised the question of whether this kind of technology puts photographers' jobs at risk. However, much of the alarm and resistance towards this kind of technology often comes from photographers who have not fully understood what AI actually is and what it actually does. The discussion on that can go way too deep into technicalities, but for the benefit of a simple discussion, let’s start by saying that AI will not mean robots will be taking over our jobs.
The video below is a discussion among myself, landscape and travel photographer Elia Locardi, and portrait and wedding photographer Jiggie Alejandrino, where we talk about how AI has affected our post-production workflow.
Artificial Intelligence in Different Forms
AI comes in various forms that have current applications in a photographer’s workflow. But from a wider perspective of technology, artificial intelligence can be seen as just a more advanced and more intricate level of computing processes that can be applied to anything and everything we do on digital devices. In a sense, the technology is not new at this point; at most, it has just matured to become more capable. AI was introduced in smartphone cameras at least seven years ago as a form of enhanced image processing through scene detection and local adjustments. This early form of AI in smartphone cameras was an improved version of the automatic mode that almost every camera has. It offered better metering and color processing through scene detection, which in effect just made it easier for smartphone photographers to take better photos without having to edit or refine so much as it uses cues to detect the scene and match applicable settings and adjustments for the detected scene. In recent months, AI-backed editing features have been made on smartphones, such as Samsung’s new S24 series, where one could easily remove objects and even subjects from the photos taken by the phone.
In conventional interchangeable lens cameras, Sony introduced AI processing in its focusing system, which first appeared on the Sony A7RV. From personal use, it was apparent that this form of AI boosts the capabilities of the processor to detect and track the selected subject through eye-AF tracking. Instead of simply detecting where the eye of the subject is and subsequently focusing on it, the AI processor is now able to predict and anticipate the motion of the moving subject to be able to track and focus faster, thereby reducing the probability of missing focus on some frames. This has been made available on other cameras that were released after it and is also now being offered by other camera brands.
AI in Post-Production
On the other hand, AI takes more diverse forms in post-production. Various tools have emerged that introduce AI processes into different steps of post-production. Some tools use AI to automate the process of developing raw images. These tools use AI-powered scene detection to apply exposure, color, and contrast treatment that is appropriate for the specific shot. Some tools use AI to detect commonly removed skin elements in the retouching process to reduce the amount of time spent editing portraits. Some tools offer exponentially greater precision in cloning out and filling in identified spots on a surface.
What needs to be said about AI in editing is that almost none of these tools are actually new. For decades now, photographers have been using tools like content-aware fill, which is just a more intuitive version of older tools such as the heal tool or the patch tool. With smarter AI capabilities, this has now evolved into the remove tool, which is just a much smarter version of the earlier tools that is more effective in preserving the details and continuity of visual elements that it is being used on.
As a landscape and architectural photographer, I personally experienced having to trace out the sky in images using tools like the pen tool, the magic wand, or the lasso tool for hours on a single image just to be able to selectively edit them or blend in a new sky. In the past couple of years, this was made easier by a smarter “select sky” option or a semi-automated sky replacement module that allowed the user to select any image and use it as the sky on the photo being worked on, and that includes images (supposedly) legally acquired to be used for that purpose. This ultimately automated the process that used to take hours and compressed it into a few seconds. Now with the use of those tools and with generative fill, the process is made much more efficient and infuses options generated from what are supposedly consensually referenced images. All of that said, all of these were technically possible with Photoshop in the past; the difference lies in the speed and precision with which the tasks are performed.
Generative AI and Photography
Perhaps the form of AI that causes much of the worry would be the generative form of AI that has the ability to create illustrations either as standalone images or as part of a photograph that is being edited. It is undeniable that there have been instances where AI-generated images were maliciously used in photography competitions. There have also been instances where commercial campaigns were called out for using AI-generated illustrations, specifically for brands that sell tools for digital artists. It is undeniable that, in the wrong hands, AI can be used to cheat, mislead, or undermine photographers and, on a larger scale, it can be used for spreading disinformation.
However, at this point, it does not seem like AI tools can do anything that human creatives cannot. If anything, they can only do them faster but not necessarily better. The core of AI is to emulate human responses to automate what was previously only doable by human work, but it seems (at least from how I see it) that the effect of what is done with AI depends entirely on the intention of whoever creates with it or whoever uses the images. That then simplifies this sophisticated artificial intelligence technology created from learned neural responses as merely a tool. In the same way that a hammer can be used both to build and to destroy, AI tools, from this perspective, remain to be what they are — tools. This is where laws, regulations, and ethics need to step in; to prevent this new tool from being used to cause the same old harm that we have been preventing in the past.
Implications of AI in Photography
The possibilities that come with AI are endless and almost impossible to bind, as with any other form of advancement in technology. It can definitely lead to unfortunate circumstances when used improperly, which is where the importance of human intervention, regulation, and government steps in. When it comes to how generative AI can be used for malicious things, it is up to us collectively to step in and prevent this.
On the other hand, AI tools in photography do offer a lot of possibilities. As with any other tool, AI can offer much better efficiency in our work as photographers. In-camera, AI has been helpful in minimizing the role of luck and increasing precision in our cameras. In post-production, AI tools can offer to do days' worth of work in only a matter of minutes if properly used and prompted to achieve the correct results. Any photographer who shoots raw can be spared from hours of raw developing or using presets that still need refining, but this can only happen if the photographer has developed a consistent style and intentionality. Portrait photographers can now more efficiently retouch their subjects based on their style, preference, and the extent of processing they choose. Commercial photographers can now do everything they’ve been doing on Photoshop for the past couple of decades with much more precision while only taking a minuscule fraction of the time.
Authenticity Will Never Lose Value
As a landscape photographer, I can say that in some way of looking at it, landscape photography was as seemingly threatening as AI is now to landscape painters and visual artists before the dawn of photography. However, there seems to be barely any conflict between landscape painters and landscape photographers because the demand for either of them is just based on the preference of an entirely different audience. At the same time, as an architectural photographer, the most constant challenge is getting a client to spend more for photography of the actual structure instead of just using the already-available design perspective that was there way before the building was erected. In this case, it doesn’t seem like AI-generated images will be any more of a threat than other ways of illustrating an architectural project.
For portrait photographers, wedding photographers, event photographers, even more so, do I personally believe that authenticity, specifically the value of having genuine images of a person or event, probably won’t decline. AI, specifically generative AI, will definitely cause and actually has caused ethical dilemmas in terms of the crisis of disinformation and misinformation, but overall, photography is the art form that has the most potential to uphold truth and authenticity even in the age of advanced technology, for as long as the line between art and documentation is clearly maintained.
But for an individual photographer, with this not-so-new advancement in technology, and with human intervention factored in, AI tools offer two options. We can either spend less time editing so that we can take in more clients and shoot more, or we can spend less time editing so that we can spend more time for ourselves outside of work. If anything, this might be a reminder that most of us probably started learning photography for the sake of taking pictures — not editing them, so technology is now offering to take the bulk of that (still) necessary step of the process so that we could focus on the creative aspect.
"...believe...probably...." Hmmm.
I think that AI generated images will pretty much eliminate pro photography.
If an individual amateur (the word comes from Latin - love/lover) photographer wants to continue, they will. I believe that won't end for at least another generation.