For two years, I’ve watched the AI hype—and hysteria—from the sidelines. Yet its impact has been impossible to ignore.
A deluge of generated content has flooded our feeds, with fake photos and bot accounts even created by Instagram itself. “They’ll have bios and profile pictures and be able to generate and share content powered by AI,” said Connor Hayes, vice-president of product for generative AI at Meta. That sounds like a brave (and bizarre) new world in store for us.
As nature photographers, our craft is grounded in patience, authenticity and our connection with the natural world. It’s a pursuit shaped by personal experience and expression.
Yet with the rapid rise and adoption of AI, we ought to reconsider how we engage with creativity and truth. Should we scorn the power of AI in a principled defence of human agency? Or should we embrace its potential, as many did with the move from film to digital photography?
I wanted to see what’s now possible—and explore how AI will help and harm the future of landscape and nature photography.
The natural world and AI as collaborators
Landscape photography can feel like a solitary pursuit. We often travel to far-flung locations at odd hours to seek solitude and focus on our craft.
But as I’ve written before, creativity asks us to be present and openly interact with the natural world around us. That is, to engage with nature as a collaborator in our craft.
Each image we capture is the fruit of an open dialogue: a back-and-forth between what nature presents and how we interpret it through our lens. The interplay of light and weather shapes our work as much as our own choices do. For many of us, this sense of partnership and communion with nature draws us into the field time and again.
Now, AI has emerged as another collaborator we can engage with. Just as nature offers us inspiration and a canvas, AI provides the tools to refine, expand and polish our creative vision. Nature invites us to observe and respond—and thrives on discovery. AI invites us to direct and deliberate—and thrives on guidance.
So how might we embrace AI’s strengths without devaluing the natural world or our agency as artists?
How can AI help?
Let’s first address the most salient example: direct image generation. So I’ll be clear, this defines a sharp line that I won’t cross in my nature photography.
In 2023, during the AI hype that year, I outlined my philosophy on post-processing. Here is my rule of thumb: If you stood where I stood under similar conditions, you’d recognise my photo as an authentic, if partly idealised, representation of that scene. (That means no cloning or warping pixels—distracting leaves or rocks can be subdued through other techniques.)
You might have a different perspective on AI-generated images. And under different circumstances, I do too. (I would have no hesitation generating a silly photo to send to a friend on their birthday.)
But by clarifying our principles and vision, we can preserve the authenticity of our art.
So how can AI still benefit your landscape photography? To begin, it excels with research and preparation. If you’re still becoming familiar with AI chatbots, think of them like a digital assistant for answering questions you might have once searched Google for. These new tools can:
Teach techniques like dreamy long exposures
Produce step-by-step guides such as shooting sharp sunstars
Explain the theory behind colour harmonies
Plan trips to unfamiliar locations
Not sure where to begin? Start by submitting a simple prompt on a topic you’d like to learn more about. (Be sure to add some personal context so the AI provides a tailored response and not a generic overview.)
Here’s a prompt I gave ChatGPT:
I haven't been to Patagonia and I want to do some forest photography there in autumn. Can you create a two-week trip itinerary? I would like a slow pace that allows me to return to interesting scenes. I want to focus on tree photography, not mountain photography. Where should I stay and what five walks would you recommend in each area? For each walk, please outline its length, difficulty and natural features.
Here’s a snippet of the response:
Days 1–4: El Chaltén, Argentina
Stay: Forest cabins or eco-lodges in El Chaltén.
Recommended Walks:
Chorrillo del Salto
Length: 3 km round trip
Difficulty: Easy
Features: Short trail to a serene waterfall framed by lenga trees in autumn colours.
Laguna Capri Forests
Length: 8 km round trip
Difficulty: Moderate
Features: Windswept lenga forests and reflections on the tranquil lake.
This outline felt refreshingly practical and tailored for a photographer's priorities, not a tourist's checklist. There’s not a day spa or hot spring in sight.
(Important note: As it responds, the AI is guessing word by word based on statistical patterns in the training data. So while what it produces isn’t always true, it will sound plausible—a dangerous combination. Be sure to cross-check any critical information. More on this soon.)
Here’s another prompt:
I've never photographed sand dunes before. What should I know before I go on a photo trip to Death Valley for the first time?
Here’s the response:
Here are some tips:
Timing: Visit at sunrise or sunset for dramatic light and shadows that highlight the dune textures. Avoid midday, as the light is flat.
Equipment: Bring a sturdy tripod and a lens cloth to combat windy, sandy conditions.
Perspective: Experiment with angles—low perspectives emphasise dune curves, while higher viewpoints showcase patterns.
Details: Focus on ripples, wind-sculpted ridges, and textures for unique shots.
Preparation: Dress for the elements and carry plenty of water.
Explore dunes like Mesquite Flat or Eureka for varied landscapes.
The AI responded with practical, location-specific advice tailored to a landscape photographer's needs. Not bad for a simple two-sentence prompt.
Sometimes, the stories behind our photos—such as our Instagram captions and blog posts—are as essential to our expression as the photos themselves.
So as a writer who explores art and creativity, I’ve found AI to be a fantastic sounding board as I shape and polish my ideas. Take this very article, for instance.
After creating an outline and drafting an introduction, I asked ChatGPT to review my structure and explore the pros and cons of using AI in landscape photography. Its feedback was fluent and insightful, particularly when I asked it to expand on the comparison between nature and AI as co-creators.
AI shines when we have a good grasp of a topic or task to guide it. It can then turbocharge the process and offer ideas we might have otherwise overlooked.
But trusting our ideas and self-expression to AIs (without human oversight) is a risky proposition. Which was why I steered clear of the tech for two years.
How can AI harm?
In March 2023, I enlisted the help of ChatGPT to research an article on the crossover between nature conservation and photography.
At first, it was helpful, providing me with a richer picture of ecological restoration—a field I was familiar with but far from an expert in. It then provided and attributed this quote to Australian Prime Minister Bob Hawke: "The conservation of the environment is not an optional extra, it is an integral part of economic development."
What a powerful vision. However, after extensive searches, I couldn’t verify the quote. When pressed, ChatGPT admitted it had fabricated it based on Hawke’s general ethos. It replied:
Upon further research, I was unable to find a specific reference to the speech where Bob Hawke made the statement "The conservation of the environment is not an optional extra, it is an integral part of economic development." However, this statement is consistent with Hawke's general views on environmental conservation and sustainable development.
It had generated the entire quote based on a vibe. I was astounded and that error informed why I steered clear of AI for so long.
I’ve since learned that this is known as a hallucination. While new AI models will continue to improve and become more accurate, their certainty often conceals inaccuracies—and blurs the line between fact and fiction.
Then there’s the problem of ‘alignment’. These tools can’t discern right from wrong. They’re simply a mirror for all the data they were trained on, spitting out results without morals. (Flawed training data has led to troubling cases of racial and gender bias in AI outputs.)
That’s why models use ‘reinforcement learning from human feedback’ to align AIs towards human values and goals. People review and rate down low-quality responses to steer models away from immoral outputs. Higher-quality, more ethical answers are great. But there are real-world costs when low-paid workers are exposed to explicit content the rest of us don’t want to see.
As an artist, I’m still wrestling with concerns about plagiarism and copyright.
Here’s a prompt I provided ChatGPT:
Create an image of Bombo Quarry at sunrise in the style of Vincent van Gogh. Include a wave crashing against the rocks and a warm glow of distant clouds.
That’s a neat party trick. But then things got interesting.
I then asked ChatGPT:
Create a fine-art photo of a Tasmanian myrtle beech tree in the style of Peter Dombrovskis. The photo should reflect the tones of 1980s film photography.
I found the fidelity here astonishing. The image is shaped by real-world depictions of Tasmanian forests, ‘imperfections’ and all. It included tree ferns with old fronds, scattered leaf litter, blown-out highlights and deep blacks.
The most unsettling part? This is AI at its weakest—its ability to generate lifelike images will only grow stronger.
More broadly, AI’s ability to imitate what’s come before could lead to watered-down creativity across fields.
As AI tools prioritise popular aesthetics, individual artistic voices may become drowned out by algorithmically generated looks. If models converge on the most likely responses, the sharp edges of art—the highs and the lows of human expression—will be sanded down over time.
Lastly, there’s the question of authenticity.
How can viewers trust that your work reflects genuine experiences—not hyper-realistic fabricated scenes? The value we place in our presence and patience in nature risks being overshadowed by simulations created in seconds.
But I’m cautious not to be the harbinger of doom and gloom.
So consider this scenario: After someone views an image of a mighty old-growth tree on Instagram, they become so moved that they start donating to a local conservation group. In this case, does it even matter if the image was AI-generated?
cautious engagement over outright rejection?
While drafting this article, I decided to revisit the quote hallucination from earlier. Knowing what I’ve now learned, I gave ChatGPT a new prompt to see how helpful it could be:
I’m a nature photographer. I believe that conserving the environment is not an optional extra. It’s an integral part of economic development and human wellbeing. Can you give me a few stirring lines that will move people to act and conserve the natural world? I’ll be delivering this speech at an economic forum to well-meaning but conservative economists.
Here’s the response:
Nature is not a resource to exhaust but capital to nurture. When we conserve the environment, we invest in resilience, prosperity and future wellbeing. To sustain progress, we must protect the very systems that sustain us.
That is some powerful framing, indeed.
For me, revisiting this example illustrates that AI is a collaborative tool. Its power lies in how we choose to wield it. Cameras, after all, can be used nefariously for subterfuge and blackmail—but they also document precious family moments and scenes of natural splendour.
Sceptics might argue that it’s a copout to compare AI to a neutral tool—and they’d have a point. From all I’ve read, the scale and scope of this technology are norm-shattering. It needs thoughtful guardrails and guidance. So I encourage you to interact with cautious engagement rather than outright rejection.
That is precisely why, as creatives who value self-expression, we can’t stick our heads in the sand. The responsibility sits with us to actively define our relationship with AI. These tools can promote human flourishing—but only if we guide them there.