There are ways to look at the images that AI is pulling from, simply by putting an image into a search over at HaveIBeenTrained.com. This is how I found that my own photography has been used to train AI, but also how I’ve been able to take an AI-created image and see what imagery was used to create it. The results are, well, photographic. And that’s really the moment we’ve been waiting for, isn’t it. A world where we can recreate a photographic-looking image without taking a photo at all. We live in that world now.
Pausing for a moment to let that sink in.
I should feel unsettled, I know. I just don’t.
I’ve spent months now working in AI and creating imagery. Most of it never goes anywhere. The process is often the same— I have an idea in my head, I go create it (which is not as easy as one might think), I get impressed by it, I save it and then… I almost never share it. Days later I look at it again and I’m simply not that impressed by it anymore. It slips into the stream of imagery I’ve made over my lifetime: drawings, designs, illustrations, photographs. Some enormous library of visuals that are nice enough, but for which only a very tiny percentage really withstand the test of time.
And this echoes my experience watching images stream through within the AI communities at large. Some constant vacillation between “wow” and “whatever.”
What happens between Wow and Whatever? How do we become so quickly inured to impressive things? This, to me, is where the discussion gets interesting. And where the true understanding of where photography is going lies.
Have you ever watched a film from the 60’s or 70’s that moved purposefully slow? Take a look, for example, at this 5-minute love scene from the film Play Misty For Me that came out in 1971.
