Looking at these comparisons, AI has done a good job at recreating Perrier’s originals - and doubtless before too long they will be even closer aligned. Subjects share a likeness, are compositionally matched, with close colour profiles. However, the lighting and texture, even after giving quite explicit prompts, aren’t so successful. The AI images almost look gloss coated, lacking the imperfections of the originals. They almost feel too contemporary; transcending the time, place and technology that produced their originals.
And across the board, the thing that AI simply cannot recreate is the humanity behind Perrier’s images. The human story that the subject carries with them, contextualising the image and (depending on your subjective taste) triggering a poignant emotional response in its viewer. Put simply, Perrier’s images are grounded in humanity.
No matter how believable the final product, even if (or indeed when) it’s indistinguishable from an ‘actual’ photograph, without a real human subject AI-produced photography cannot have the same depth of storytelling, the same depth of meaning. And sure, we could generate an AI-fuelled story to accompany the image - but the fact (or lack thereof) remains the same.
AI might be the latest in a long history of innovations to challenge creativity as we know it. But powerful and intriguing though it may be, its automation means it cannot exist in the same category as creative innovations - like photography - that have come before it.
Only time will tell whether we will ever respond as emotionally - either positively or negatively - to an AI photography exhibition as we did to Perrier’s work. But there’s no doubt that AI will change, is already changing, what we understand the concept and parameters of art to be.