Closer #010
This weekly experiment is on an indefinite hold, until significant AI-advancements have been made.
Curious about the conclusions? Read Photography in the Age of AI (2025): Lessons From 25 Weeks of Experimentation
Welcome to Closer,
A weekly experiment in AI interpretation: I share a photo, ask AI to describe it, then generate an image from that description. This week we're skipping the description part though, for the sake of experimentation.
Human experience → photo → AI image. Edition #010.
Prompts are in grey boxes. Descriptions are in italics.
LLMs used
GPT-4o for image interpretation and image generation.
The input
The image above is part of my black and white square-format photography project made during the winter of 2020/2021. Take a look at the project page to find out more.
The output
Impressions
My thoughts from last week are underlined by this week's output. This might be the most accurate replication of the input image so far. Granted, we've also skipped a step and therefore significantly reduced the chances of things getting lost in translation. That direct connection between the real world and AI trying to synthesize it makes it clear that life can ben replicated. At least to a level where it's good enough for some use cases. But I would argue that's hardly a surprise at this stage.
What I'd like to highlight though, is how the synthetic images continue to lack any indication of soul. The imperfections of human-created life I talked about last week are completely stripped from the image. Which is striking, when you consider we've specifically asked the AI to replicate the source image. Why do we still end up with an output that's a little more smooth, a little less detailed, feels so much less real?
Perhaps a preliminary conclusion for AI at this stage, is that it cam mimic the subjects and what they might look like in a perfect world. But our world isn't perfect. The weather wears the wood on the fencing down. The sand of the beach is shaped by the wind in a flowing manner, but continues to display a coarser texture than any AI is willing to do right now. The quality of the light in real life is less flat, has more depth, more ✨spark✨. But is that spark uniquely ours? If so, will it continue to be?
Anything else that caught your eye? Feel free to send a reply. Please also let me know if you have questions, suggestions or remarks about my approach. This is just as much a journey for me as it is for you.
See you next week.
Mitch
Just you, me, and some occasional notes from the field. No spam.
Join the conversation