How AI Is Changing the Landscape for Stock Photographers — and 5 Ways I’m Adapting

Published by

on

I’ve been submitting images to stock agencies since 2015, and lately, the landscape feels different — almost unrecognizable. Artificial Intelligence has swept in like a quiet storm, filling the market with endless images that never existed: perfect models, ideal lighting, flawless compositions. Buyers can now create “a smiling team in an office” or “a woman drinking coffee at sunrise” in seconds, for almost nothing.

That shift has made it clear: generic stock is in trouble.

In 2023 I made a prediction, that this would happen. The kinds of images that used to sell steadily — polished business scenes, clean-cut lifestyle shots, cheerful “diversity” portraits — are now easily replaced by AI. You can feel it in the sales numbers: those safe, conceptual photos that once formed the backbone of stock portfolios just aren’t moving like they used to.

But not everything is vanishing. If anything, this change has reminded me what makes photography human. What’s holding strong — even growing — are the things AI can’t replicate convincingly: real places, real people, and real stories.

Editorial and documentary-style images, grounded in identifiable locations or authentic moments, are more in demand than ever. A real street scene in Lisbon, a local artisan at work, a foggy morning in a recognizable park — those are things algorithms can’t fake without tripping over geography, ethics, or copyright. Editors and buyers still need genuine visuals to illustrate real-world stories.

Here’s what I’ve learned about staying relevant — and how photographers like us can keep selling in this new era:


1. Photograph the Real World

I’ve shifted my focus toward editorial and travel images — things that exist. When you capture a recognizable place, an event, or a slice of life, you’re creating something grounded. AI can imagine a cathedral, but not that cathedral.

The FEATURED IMAGE, one my best selling images across agencies, is the landscape of a local park (Goat Rock State Park).

Below, a few more example of good earners:

2. Work in Niches AI Can’t Reach

AI doesn’t have access to your world — your neighborhood, your community, your ecosystems. I’ve started documenting local people at work, conservation projects, and regional traditions. These are things only a human being, physically present, can photograph.

3. Tell Small Stories

I try to shoot in sequences — images that suggest a process or story rather than isolated objects. AI can produce individual pictures, but it can’t create narrative flow. A baker kneading dough, then pulling bread from the oven, then sharing it with a customer — that’s storytelling, and buyers love it.

4. Be Precise With Captions and Keywords

With so many images online, context matters. I include specific details: who, where, when, and why it matters. “Fish market in Puntarenas, Costa Rica, early morning” performs far better than “people at market.” Editors are searching for real, nameable places.

5. Don’t Rely Solely on Stock Agencies

Agencies are still useful, but I’ve been branching out — selling prints, licensing directly to NGOs and small publishers, even sharing behind-the-scenes stories on my website. Building visibility beyond agency algorithms gives you more control and stability.


The way I see it, AI has taken over the fictional side of photography — but it can’t touch the truthful side. There’s still value in seeing the world through a real lens, with all its imperfections, timing, and emotion.

Stock photography is evolving, but so can we. The photographers who will thrive are the ones who double down on what’s real — because in a world full of synthetic images, authenticity has never looked so good.

Here are links to some of my portfolios:

Shutterstock

Adobe Stock

DreamsTime

Alamy

______________________________

Wall Art Botanical Images

Wall Art Photography projects

Wall Art landscapes and miscellaneous

________________________________

10 responses to “How AI Is Changing the Landscape for Stock Photographers — and 5 Ways I’m Adapting”

  1. alneyman Avatar
    alneyman

    Alessa

    Wonderful piece – your undaunted attitude is inspiring. Looking forward to “your next” – please keep it up!

    Al Neyman

    Like

    1. Alessandra Chaves Avatar

      Thanks, Al

      Like

  2. shoreacres Avatar

    I like this, and agree with it: “The way I see it, AI has taken over the fictional side of photography — but it can’t touch the truthful side. There’s still value in seeing the world through a real lens, with all its imperfections, timing, and emotion.”

    I’ve read similar observations here and there. You’re not alone.

    Liked by 1 person

    1. Alessandra Chaves Avatar

      I have noticed a sharp increase in AI-generated videos on the Internet, usually broadcasting fake news.
      It’s sad and I don’t know how a person is supposed to protect themselves from this.

      Like

  3. Steve Schwartzman Avatar

    Have you experimented with asking AI for a photograph of a specific place like Goat Rock State Park to see how accurately (or inaccurately, or not at at all) it creates an image of that place?

    Like

    1. Alessandra Chaves Avatar

      I have. I asked AI why it doesn’t generate accurate images of places. Here is the answer:
      1. AI can’t access real photos or maps
      Image-generation models don’t “look up” real locations (like Google Maps or real photos).
      They create images based on patterns learned during training. So when you say “Goat Rock State Park”, the model tries to recreate the general look of a Northern California beach with sea stacks — but it cannot reproduce the exact, real-world geography.
      2. AI models often mix up landmarks
      Goat Rock has a very specific shape and sits at the mouth of the Russian River, with:
      the river lagoon area
      the road and parking lot above the beach
      Jenner visible nearby
      the very distinctive flat-topped Goat Rock itself
      Models sometimes substitute generic sea stacks that look “PNW/Oregon Coast,” because those shapes are more common in their training data.
      3. It generates “plausible,” not “correct,” scenery
      Even when asked for a precise place, the model mainly aims for something believable rather than geographically accurate.
      So it may invent:
      incorrect coastline shape
      wrong rock formations
      hills or cliffs that don’t exist
      lighting and angles that aren’t real
      4. Prompts like “State Park” trigger generic natural-park aesthetics
      This sometimes pushes the model toward more idealized or dramatic landscapes.

      Like

      1. Steve Schwartzman Avatar

        It’s edifying to see how far you’ve delved into the subject. Have you considered writing an article about that and submitting it to a magazine or website devoted to either photography or AI?

        Like

      2. Alessandra Chaves Avatar

        That is actually a good idea!

        Like

  4. tierneycreates: a fusion of textiles and smiles Avatar

    Proactive and brilliant and AI doesn’t have your eye!

    Like

    1. Alessandra Chaves Avatar

      AI’s eye… 😉

      Liked by 1 person

Leave a reply to Steve Schwartzman Cancel reply

Discover more from It is all about the light

Subscribe now to keep reading and get access to the full archive.

Continue reading