In April 2023, I declined the Sony World Photography Award because the image I had submitted was AI-generated. My refusal of the prize was motivated by the need to draw a clearer distinction between photography and AI-generated images (promptography), and to spark a discussion about the future relationship between the two. This essay aims to contribute to that discussion by highlighting the differences in motivation and working methods, and by taking a look at the future collaboration between humans and machines. THE WORLD OF THE PHOTOGRAPHERPhotography is the art of creating images with light. For most photographers, this means “going out into the world.” Many photographers are driven by curiosity about people, cultures, and places. Their hunger for experiences is fueled by empathy. Photographers want to freeze time, build bridges to other ways of life, raise awareness of problems, give a voice to the voiceless, and show the “human condition.” The image is created through a dance with the world—a collaborative process with whatever is in front of the lens. When working with people, respect, trust, and empathy are the keys to powerful images. Social competence pays off; in a good portrait, it’s about finding and being found. PROMPTOGRAPHY & THE WORLDDoes promptography require a “journey out into the world”? No. But promptographers must have been out in the world at least once to gather experiences and knowledge—because that is the material they work with. In prompting, the journey is ideally inward. I must ask myself: What do I want to create? And why? I can draw entirely from my imagination. But what comes of that? What experiences can I build upon? However, many people do not embark on this potential inward journey. A closer look at user categories explains why. The “Average Users”Most users welcome AI as a shortcut to an image result they couldn’t previously achieve with their own skills. They lacked time, talent, training, or patience. The speed and ease with which they can now produce images that surpass their previous abilities excites them. They are not interested in depicting the “human condition” or freezing time. What matters is creating a cool image that gets applause from their community. It doesn’t matter whether the user was the driving force behind the image or was simply carried along by the AI. The result is what counts. And the result often reveals the unconscious clichés that preoccupy these people. Here, we find many mashups, such as Batman dressed as Barbie. The “Prompt Engineers”Professional image creators use the new technology to save time and money. They want to stay competitive and up to date. Prompt engineers are experts in a specific image category and want to solve a particular task with AI. They know the keywords and workflows of an industry and already have experience in it. To use AI image generators as a professional tool, I must enter the process with the intention of solving a task. The first generated images are just the beginning—the starting point for continuous fine-tuning, where interim results are critically examined, discarded, regenerated, workflows adjusted, and platforms remixed. To evaluate the generated output and make corrections, I need experience in image design and editing. Prompt engineers don’t just give commands; they give informed instructions based on knowledge and experience in their field. The “Prompt Whisperers”As an artist, I see the technology as a liberation of individual imagination from material constraints. Since no compromises need to be made in production, and the subject, action, location, lighting, equipment, etc., can all be freely chosen, the artist can, for the first time, express their vision independently of material limitations. The promise of being able to create new visual languages with a new tool—synthesizing unprecedented aesthetics from the visual languages of the past—is enticing. Unlike prompt engineering, it’s not about creating a product or serving a target audience. Like many photographers, prompt whisperers are interested in expressing the “human condition.” Since the AI training data can be seen as a mirror of humanity—or what C. G. Jung called the “collective unconscious”—the task of the prompt whisperer is to synchronize their own subjectivity with the collective unconscious. In doing so, they become the conductor of the anonymous choir formed by the AI’s training data. To become a prompt whisperer, you need the skills of a prompt engineer, an artistic mind, and a deeper awareness of your own personality. If you fail here, your work will reveal the unconscious clichés of humanity. If you cannot optimize and conduct the training data, the training data will take over and produce a statistical average.
|