Photographer Oh Seung-hwan: AI Can’t Match Human Vision in Photography

The Evolution of Photography in the Age of AI In an era where artificial intelligence (AI) is reshaping the way we create and perceive images, the photography world stands at a crossroads. Since the rise of ChatGPT, a multitude of generative AI tools have flooded the market, transforming initially crude outputs into increasingly sophisticated visuals. […]

The Evolution of Photography in the Age of AI

In an era where artificial intelligence (AI) is reshaping the way we create and perceive images, the photography world stands at a crossroads. Since the rise of ChatGPT, a multitude of generative AI tools have flooded the market, transforming initially crude outputs into increasingly sophisticated visuals. The rapid pace of AI development has sparked both excitement and concern, with many questioning whether AI could eventually replace traditional photography. Amid this uncertainty, one photographer has boldly declared that AI images represent a revolutionary shift in the field.

Oh Seung-hwan, a seasoned photographer who studied and taught photography for over two decades, recently released a book on AI photography and held an exhibition at the Insadong Gallery Index in Seoul. His work, displayed using Midjourney, showcases AI-generated images that closely resemble traditional photographs. These works are accompanied by edition numbers and price tags, highlighting the effort and artistic intent behind each piece.

Is Photography Over?

Oh Seung-hwan asserts that the traditional form of photography is no longer the same. He explains that the term “post-photography” has been discussed in the global art world for over 20 years. This concept suggests that the act of creating photos with cameras may soon be replaced by AI-driven image generation. According to Oh, the process of generating images with AI is similar to the earliest human attempts at visual representation, such as cave paintings. It’s not about capturing what is visible but about visualizing thoughts and ideas.

He emphasizes that photography has never been solely about recording reality. Even documentary or news photography involves subjective choices in framing and composition. The final image is determined by the photographer’s decisions, making it a creative process rather than a mere mechanical act. As AI becomes more prevalent, the definition of photography itself may need to evolve.

Is the Camera Obsolete?

While AI can produce realistic images, Oh argues that the camera still holds a unique role. The ability to capture moments—specific instances of time and space—is something only humans can do. AI lacks the capacity to understand human expressions, emotions, and context. It can generate content, but not the deeper meaning behind it.

Photographers will continue to play a vital role in interpreting and reproducing landscapes, capturing the essence of a scene, and conveying the emotional depth of a subject. While AI can create images, it cannot replicate the human experience of observing and recording the world. Cameras will continue to serve as tools for proving existence, capturing the experiential values of light, air, and touch.

AI Photography: The Ultimate Digital Retouching

Oh shares how he used AI to recreate a scene from 40 years ago—a sheep flock in front of Mont-Saint-Michel. By inputting a photo he had taken, AI generated the desired image within 40 minutes. This process mirrors the evolution of digital editing, which once faced skepticism when Photoshop first emerged. Initially, many photographers dismissed digital manipulation as not “real” photography. However, as technology advanced, the line between original and edited images blurred.

AI does not create fakes; it is the user who determines the outcome. The quality of AI-generated images often reflects the data it was trained on, leading to biases such as overrepresentation of certain demographics or cultural elements. To achieve the desired result, prompts must be carefully crafted and refined through multiple iterations.

AI Photography: Deceptively Simple, Labor-Intensive Work

Creating AI-generated images requires more than just a few keywords. Oh explains that without a clear mental image, the process is impossible. For example, describing a child running on a beach involves specifying details like clothing, movements, and expressions. His prompts are extensive, including camera models, shooting times, and angles.

Generating a single image often involves producing hundreds of variations, refining them until the desired result emerges. Each image can take hours of work, with some requiring up to ten hours of adjustments. The process is both challenging and rewarding, reminiscent of the slow reveal of a photograph in a darkroom.

What Lies Ahead for Photographers?

Oh believes that photographers and AI will coexist. News and documentary photography will remain human-driven, while commercial and artistic photography may see increased use of AI. Staged or graphic-heavy images are likely to be replaced by AI-generated alternatives. However, as AI data expands, current biases and imperfections will improve over time.

The exhibition featuring Oh Seung-hwan’s AI photography runs until the 15th, offering a glimpse into the future of visual storytelling.