Artificial intelligence sees the world very differently from you and me. It doesn’t feel awe when it sees mountains. It doesn’t recognize poetry in someone’s tired eyes on the subway. But it does see — just not like we do.
So what happens when we ask a machine to look at the world through a lens? And perhaps more interestingly: what can that machine’s way of seeing teach us about our own?
This isn’t a post about replacing photographers with AI (or vice versa). It’s about something more curious. What can machine vision reveal about the biases, intentions, and instincts we bring to photography as humans?
Let’s explore.
How Machines “See” Images
Most AI image systems — like facial recognition software or surveillance cameras — aren’t designed for people to interpret. They’re designed for machines to understand.
They break images down into data: pixels, shapes, contrast, depth, heat. They learn through training sets, millions of images labeled with meaning. The extracted patterns help them detect, identify, or classify what they “see.”
This kind of seeing is purely functional. It’s not about beauty or meaning. But that stripped-down, hyper-objective vision forces an interesting question:
What would your photography look like to a machine? Would it detect light or shadow first? Would it recognize your subject at all? What story would it infer from your choices — color, framing, angle?
And: what does your human perspective add that the machine can’t?
What Photographers Can Learn from Machine Vision
Let’s go deeper. Here are a few questions AI systems force us to consider as photographers:
1. What do you assume is meaningful in an image?
AI doesn’t care about emotion unless it’s trained to detect expressions. It doesn’t know a protest sign from a streetlight — unless it’s told which is which.
This raises a provocative reflection: are we taking photos based on emotional instinct, or learned visual cues? Have we trained ourselves to seek certain types of “meaning” without questioning it?
Try this:
Take a photo and write down everything an AI might see: shapes, people, lines, objects. Then write what you saw — feeling, metaphor, story. Notice the gap between machine seeing and human storytelling.
2. How much do you rely on visual tropes?
Machine vision often reinforces existing structures, like tagging women smiling as “friendly,” or dark scenes as “dangerous.” These come from the biases in the datasets they’re trained on.
Now flip the lens: how much of your own photography is shaped by visual clichés or societal bias? Do you associate wide angles with adventure? Black and white with seriousness? Sharpness with truth?
Try this:
Shoot a scene intentionally outside its visual trope. Make joy look somber. Make solitude look crowded. Break the expected signal.
3. What does your framing leave out?
AI sees everything in its scope — but what it recognizes depends on its parameters. A photo of a peaceful protest could be flagged as “suspicious activity” based on posture or density.
Photography, too, is selective. Framing is an act of power: choosing what’s seen and what’s not. Thinking like a machine reveals how much context matters, and how easily images can be misread without it.
Try this:
Photograph the same scene three times with different framings: close-up, medium, and wide. What truths do each capture, or obscure?
The World AI Helps Sculpt
Trevor Paglen asks: what kind of world do these imaging systems help sculpt? When we build machines to surveil, to identify, to categorize — what visual culture do we participate in?
As photographers, we’re part of that world-shaping, too.
Every image has consequences. From street life to abstract textures, your vision — unlike a machine’s — is shaped by memory, feeling, culture.
That’s not a weakness. That’s your edge.
Final Thought: You’re Not Just a Sensor
AI is a mirror — and a cautionary tale. It reminds us that seeing is never neutral. That images always reflect more than just light. That framing, timing, and subject are acts of authorship. You are not just a sensor collecting data. You are a maker of meaning.
So the next time you lift your camera, ask yourself: What story am I telling? and What kind of world is this image helping to build?
MELISSA MARTINSEN 