Metadata & the Machines |
How AI Is Rewriting Some of the Rules for Wildlife Photographers
The jungle is restless in the pre-dawn hours. I can feel the humidity pressing in, wrapping around me like a damp second skin as I tread carefully along a narrow trail in the lowland rainforest of Panama. My boots sink slightly into the rich, dark earth, the sounds of the jungle amplified in the stillness—frogs calling from hidden pools, the rhythmic hum of cicadas just getting started for the day, the occasional rustling of leaves as something unseen moves through the understory. It’s difficult not to let the mind wander through this Heart of Darkness like moment. Jaguars. Pumas. Fer-de-lance. Bushmasters. Oh my.
This is my third morning in this forest, and I’m here for one reason: Geoffroy’s tamarins.
I pause, scanning the canopy with my field glasses. These tiny primates, endemic to Panama, where they fetch top dollar on the black-market trade for exotic and illegal pets, are notoriously difficult to find. Weighing less than a pound, they dart between branches with an agility that makes following them through the forest a near-impossibility at times. But I know they are most active in the early morning hours, traveling in small family groups as they search for food, and most importantly, they both emulate and respond to the calls of tyrant flycatchers. Where there is one, there is often the other.
A rustling above catches my attention. A flash of black and white among the green. I raise my camera instinctively, tracking the movement. A moment later, a tamarin scampers onto a lichen-covered branch, its black-and-white face framed by an unruly crest of fur.
The Hidden Work Behind Every Image
Back at the lodge, I review my photos. The morning was a success. But my work is just beginning. Muddy boots and the pursuit of adventure is only the first part of the process. Now the real work begins. Before I can send these images off to a stock agency, magazine editor, or conservation organization, I have a long and tedious chore ahead of me: metadata.
Metadata, the titles, captions, and keywords that describe an image, is the backbone of making a living with wildlife photography. Without it, even the best images become just another file on a hard drive, effectively invisible to researchers, editors, or buyers who might need them. The right metadata ensures that when someone searches for “endangered Panamanian primate” or “New World monkey species,” my image of the Geoffroy’s tamarin magically appears before them.
The problem, however, is that creating the necessary metadata for all these photos is a lesson in tedium. It takes effort. Consideration. Strategy. And a whole lot of time. A single image requires 30-50 keywords, a long-form and engaging caption, and a title that balances description with searchability. Multiply that by hundreds if not thousands of images that are kept from a single expedition, and suddenly, the process becomes overwhelming. And this why AI is starting to play a major role in automating metadata generation for me; turning what was once a painstaking task into something faster, more efficient, and, in some cases, even more accurate part of my workflow.
To even suggest that I am beginning to embrace artificial intelligence as a working photographer feels downright sacrilegious. Declaring such is likely akin to pulling the pin on a grenade and rolling it across a room between me and many other photographers. AI is the Pandora’s Box of our modern world, threatening to cannibalize all forms of creative work as we know it.
And yet, us wildlife photographers have been the beneficiary of artificial intelligence for years whether we realize it or not. Maybe it’s the Deep Learning chips in our flagship cameras that find, track, and predict the movements of our subjects, or perhaps it’s the species recognition capabilities of software such as Merlin that’s used to help identify birds. Or it could be the photography community’s newfound love affair with image enhancing software such as Topaz. Whatever the gateway drug of choice may be for wildlife photographers, AI has been an integral part of our craft for sometime now.
Personally, I have always found more common ground with the Luddite movement of the Industrial Revolution than I have with the technophiles of the digital one. I prefer to hold and smell and feel and read books made from paper instead iPads and e-readers. I hate social media, wading through its toxic sludge only when I must. I cherish the moments in life when I find myself without cell phone reception and laugh when people can’t understand why I don’t even know where my phone is most of the time.
And yet, here I sit hammering out these words on my MacBook, writing about digital wildlife photography. This afternoon, I will begin editing images I made yesterday of the most important mangrove forest in Central America with my DJI Mavic 3 Pro drone. And it will all take place while perched in a tree house of sorts overlooking the tropical cloud forest of Western Panama, connected to the rest of the known universe via Starlink. While I would rather live out my Hemmingway and Hunter S. Thompson fantasies in the tropics with an old typewriter and bottle of rum like they did, those days are gone and that would be bad for my liver.
The thing I love most about wildlife photography, and my style of nature writing, is that despite how dependent it all is on cutting edge technology and computers, it also requires me to have one foot firmly grounded in the analogue world of nature.
There was a time when all of us humans made our living by understanding the most nuanced aspects of the natural world. Today, however, very few outside hunter-gatherer societies understand how to track an animal, or that a particular type of tree is used medicinally by coatimundis in the rainforest, or that a tiny species of New World primate has something of a symbiotic relationship with tyrant flycatchers and following the sound of those birds can lead you to a Geoffroy’s tamarin. But for me, this is both occupation and a way of life. Wildlife photography allows me to live within both the digital dystopia and the rich tapestry of life outside of it, simultaneously.
As life becomes ever more entangled with technology, I find that I must balance this with equal or greater parts nature. This is what led me to finally embrace and begin using artificial intelligence in my workflow as a photographer. As I mentioned above, what happens in the field is only the first step. The real work of making a living happens not behind a lens today, but behind a computer.
For most working professionals, the photography bit of being a photographer is only about 10% of the job. The rest is trying to figure out how to make money with those photographs once you return home. Of all the things I dread the most about this part of the job, the thing that always stands as the narrowest bottleneck in the process for me, is spending days or even weeks after an extended time afield writing long-form captions, titles, and coming up with 30-50 strategic keywords for each of those images. This is time I could be writing articles for PhotoWILD Magazine, for instance.. This is time I could be marketing those images. And more importantly to me, this is time I could be in the forest doing the thing I actually love.
Geoffroy's Tamarin Calling From Rainforest Canopy
Perched high on a palm frond, a Geoffroy’s tamarin (Saguinus geoffroyi) vocalizes loudly to assert its territory in the tropical lowlands of central Panama. Endemic to Panama and a narrow stretch of northwestern Colombia, this diminutive primate is one of the few New World monkeys found only within this limited range. Its striking appearance—marked by a dark face, white crest, and agile movements—has unfortunately made it a frequent target of the illegal pet trade. In Panama, Geoffroy’s tamarins are often captured and sold on the black market, where their small size and expressive faces make them tragically desirable. Combined with accelerating habitat fragmentation, this exploitation places growing pressure on wild populations. Conservation efforts now focus on enforcing wildlife trade laws and protecting the remaining patches of lowland and secondary forest that these endemic primates depend upon for survival.
KEYWORDS
animal behavior, animal call, animal wildlife, animals in the wild, arboreal animal, biodiversity, dense forest, environment, environmental conservation, environmental problems, green background, lush foliage, monkey, natural condition, nature, primate, rainforest, tamarin monkey, territorial animal, threatened species, tropical climate, tropical rainforest, tropical tree, warning behavior, wildlife observation, no people, outdoors, animal, forest, mammal, cute, one animal, close-up, small, geoffroy's tamari, wildlife conservation, black market, illegal pet trade, Panama Canal, Soberanía National Park, deforestation
Geoffroy's Tamarin Calling From Rainforest Canopy
Caption
Perched high on a palm frond, a Geoffroy’s tamarin (Saguinus geoffroyi) vocalizes loudly to assert its territory in the tropical lowlands of central Panama. Endemic to Panama and a narrow stretch of northwestern Colombia, this diminutive primate is one of the few New World monkeys found only within this limited range. Its striking appearance—marked by a dark face, white crest, and agile movements—has unfortunately made it a frequent target of the illegal pet trade. In Panama, Geoffroy’s tamarins are often captured and sold on the black market, where their small size and expressive faces make them tragically desirable. Combined with accelerating habitat fragmentation, this exploitation places growing pressure on wild populations. Conservation efforts now focus on enforcing wildlife trade laws and protecting the remaining patches of lowland and secondary forest that these endemic primates depend upon for survival.
KEYWORDS
animal behavior, animal call, animal wildlife, animals in the wild, arboreal animal, biodiversity, dense forest, environment, environmental conservation, environmental problems, green background, lush foliage, monkey, natural condition, nature, primate, rainforest, tamarin monkey, territorial animal, threatened species, tropical climate, tropical rainforest, tropical tree, warning behavior, wildlife observation, no people, outdoors, animal, forest, mammal, cute, one animal, close-up, small, geoffroy's tamari, wildlife conservation, black market, illegal pet trade, Panama Canal, Soberanía National Park, deforestation
Automated Captioning and Keywording: How AI Enhances Image Metadata
Traditionally, the time-consuming aspects of metadata has been a distinctly manual and obnoxiously meticulous process requiring photographers to identify the subject, describe what’s happening in each image, and craft a compelling yet concise caption that balances scientific accuracy with engaging storytelling. Think of it as writing an entire article about a topic in three or four sentences. Writers have a saying that goes something like, “I would have made it shorter, but I didn’t have enough time.”
This, in many ways, has always been the challenge of cold caption writing. Making things concise, to the point, and short, takes time and planning and rewrites. When staring at a thousand images to do this with however, that’s enough to make most people throw their hands up and call a single sentence “good enough.”
Much like any sort of professional writing, you must first decide what audience you want to appeal to. With captioning a photograph of a Geoffroy’s tamarin, for instance, it could be used for eco-tourism and travel. But the same photograph could be used for academic purposes given that this species is an endemic. Another possibility is that it could be used to help educate and illustrate the horrific impacts of the illegal pet trade coming out of the tropics, or the importance of protecting biodiversity hotspots, or why the management of the Panama Canal Zone has been dependent on the creation of a series of national parks and intact forests to maintain the watershed necessary for the Canal to function.
So, which is it?
Travel and tourism?
Academia?
Illegal pet trade?
The Panama Canal?
You see, a photo of a Geoffroy’s tamarin is rarely going to sell itself. Sure, there may be a photo buyer out there specifically looking for photos of this little primate. But more often, photo buyers don’t necessarily know what they want in a wildlife photo.
It’s been my experience in the industry that photo buyers start with a concept and then purchase photos that help illustrate that concept. That is to say, an editor looking for photos to illustrate a story about the importance of national parks to the functioning of the Panama Canal may be looking for just about anything cool, unique, or compelling from the natural world that pertains to that region. This could be a toucan. It could be a drone photo of the gallery forest along the banks of the Chagres River. It could be a caiman, red-lored parrot, indigenous fisherman in a dugout canoe, or a landscape photo showing intact rainforest with the skyline of Panama City in the background. It could be all these things or none of them. Or perhaps, it could be my Geoffroy’s tamarin.
This is why metadata is a critically important component of wildlife photography. Metadata is how editors and art directors come to find your photographs in association with whatever it is they are looking to illustrate. If I simply title, caption, and keyword my image in a way that only describes a Geoffroy’s tamarin sitting on a branch in the rainforest, this photo will likely never see the light of day.
But understanding this is just the beginning. Next comes deciding how you want, or hope, your photograph to be used. A caption focused on selling an image of the tamarin to a travel agency is going to be written very different from one about the illegal pet trade in the area. You must decide, then write a compelling caption that describes the photo and puts it all into the context of travel and tourism or the black market for exotic pets.
Now do this again, over and over and over for each photo you choose to keep from a day’s shoot.
Additionally, the photographer must determine a comprehensive set of keywords; terms that will make the image searchable within stock agencies, conservation databases, or editorial archives. This task is also overwhelming, especially for photographers returning from expeditions with thousands of images.
As you likely begin to see here, the need to create accurate, detailed, and searchable metadata for each file can take weeks sometimes. And it’s here where artificial intelligence is making a transformative impact, allowing photographers like me to spend less time on “administrative tasks” and more time in the field.
White-Faced Capuchin in Soberanía National Park, Panama
CAPTIONS
A Panamanian white-faced capuchin (Cebus imitator) peers out from the dense rainforest canopy of Soberanía National Park, just outside Panama City. With its expressive face and curious nature, this intelligent primate is a favorite among visitors exploring the park’s famed Pipeline Road, a hotspot for both birdwatchers and wildlife photographers. Soberanía offers one of the most accessible windows into lowland tropical rainforest in Central America, drawing eco-tourists from around the world to its well-maintained trails and remarkable biodiversity. White-faced capuchins here are often seen foraging in social groups, leaping effortlessly through the trees as they search for fruit, insects, and small vertebrates. As one of Panama’s most important protected areas for ecotourism and environmental education, Soberanía National Park plays a vital role in connecting people with nature while safeguarding the country’s rich biological heritage.
KEYWORDS
central america, animal, animal behavior, animal wildlife, animals in the wild, arboreal animal, biodiversity, capuchin monkey, close-up, dense forest, environment, lush foliage, mammal, national park, nature, one animal, panama city - Panama, primate, rainforest, threatened species, tropical rainforest, wildlife reserve, no people, tropical climate, tropical tree, travel destinations, white-faced capuchin monkey, monkey, forest, portrait, wildlife conservation, deforestation, Panama Canal, tourism, travel
White-Faced Capuchin in Soberanía National Park, Panama
CAPTIONS
A Panamanian white-faced capuchin (Cebus imitator) peers out from the dense rainforest canopy of Soberanía National Park, just outside Panama City. With its expressive face and curious nature, this intelligent primate is a favorite among visitors exploring the park’s famed Pipeline Road, a hotspot for both birdwatchers and wildlife photographers. Soberanía offers one of the most accessible windows into lowland tropical rainforest in Central America, drawing eco-tourists from around the world to its well-maintained trails and remarkable biodiversity. White-faced capuchins here are often seen foraging in social groups, leaping effortlessly through the trees as they search for fruit, insects, and small vertebrates. As one of Panama’s most important protected areas for ecotourism and environmental education, Soberanía National Park plays a vital role in connecting people with nature while safeguarding the country’s rich biological heritage.
KEYWORDS
central america, animal, animal behavior, animal wildlife, animals in the wild, arboreal animal, biodiversity, capuchin monkey, close-up, dense forest, environment, lush foliage, mammal, national park, nature, one animal, panama city - Panama, primate, rainforest, threatened species, tropical rainforest, wildlife reserve, no people, tropical climate, tropical tree, travel destinations, white-faced capuchin monkey, monkey, forest, portrait, wildlife conservation, deforestation, Panama Canal, tourism, travel
How AI Generates Captions for Wildlife Photography
AI-powered natural language processing (NLP) tools are now capable of analyzing an image and automatically generating relevant and structured captions that describe its content. These systems rely on a combination of computer vision technology and pre-trained language models to assess the image, recognize the species, detect environmental features, and construct a readable, human-like caption.
For example, if I upload an image of a Geoffroy’s tamarin to an AI-powered metadata tool, the system may generate a caption like:
“A Geoffroy’s tamarin (Saguinus geoffroyi) perches on a moss-covered branch in Panama’s lowland rainforest. This small primate, endemic to Central America, relies on its long tail for balance as it navigates the dense canopy in search of fruit, insects, and tree sap.”
This caption not only describes what is happening in the image but also includes scientific context, a brief behavioral description, and geographical information that makes the image more valuable to researchers, stock agencies, and buyers.
But this is just the tip of the iceberg.
All of this can be tweaked and refined or expanded based around a simple set of instructions.
If we take the same photograph and add instructions such as, “create a caption for a photograph of a Geoffroy’s tamarin that focuses on the uniqueness of the species and the threat of the illegal pet trade,” we may get the following:
“A Geoffroy’s tamarin (Saguinus geoffroyi) perches on a mossy branch in the tropical forests of Panama, its striking black-and-white facial markings and golden-hued back blending into the dappled light of the canopy. As one of the smallest primates in Central America, this species plays a vital role in seed dispersal, helping to maintain the health of its fragile ecosystem. However, despite its ecological importance, the Geoffroy’s tamarin faces increasing threats from habitat destruction and the illegal pet trade. Many of these charismatic monkeys are captured and sold into the exotic pet market, where they suffer from inadequate care for the entertainment of those who can afford to pay for smuggled wildlife. Conservation efforts are critical to ensuring the survival of this species in the wild, highlighting the urgent need to protect Panama’s remaining forests and combat wildlife trafficking.”
Now THAT is a caption.
It’s full of SEO (search engine optimization) for a photo buyer looking to illustrate an article or conservation campaign about the illegal pet trade. It educates the photo buyer as to why this photo is the one that needs to be purchased for their project. And it’s a heck of lot better than “a monkey on a branch in Panama.”
All of this was once manually entered into the metadata field in Lightroom. Even a handful of images could easily take an hour or more to build out captions for like this. But AI can now do this tedious chore in a matter of seconds.
One of the significant advantages of AI-generated captions is just the consistency in format and style. When manually tagging images, photographers often struggle with variations in wording, phrasing, or terminology, which can affect searchability. We get tired of doing, have creative blocks, get frustrated, and after 8 hours of doing it I’m usually quite brain dead.
Whether you are simply posting photos on your website or uploading them to a stock photography agency, Google crawls the associated metadata of the file you upload, indexing the image accordingly and showing the photograph in relevant searches. AI ensures that captions remain uniform, following a structured format that aligns with the requirements of stock agencies, conservation organizations, scientific archives, and even how Google compiles data.
The above photo is a screenshot of Pixify, configured specifically for Getty's Controlled Vocabulary. Thus far, this is the best program I have found for generating keywords for use with Getty Images. While this still requires significant oversight (Pixify tagged on of these photos as "green-golf course"), it gets pretty close.
My experience has been that Pixify doesn't like a wide diversity of different types of images all at once. Applying it to a series of Talamanca hummingbirds, for instance, tends to produce better results than hummingbirds, monkeys, and drone photos all thrown in together.
Some of this has to do with the way in which Pixify allows you to add instruction. On the right hand side, you can see "file instructions." Here you can in and add instructions to the program, telling it what the photos are of and how you would like to approach keywording. But above that, not seen in this photo, is a box for adding global instructions for all of the photos. And it's here, when I give Pixify direction about the series of photos, that I find everything to be the most accurate.
Keywords for Improved Searchability
Keywords are arguably the most important part of metadata when it comes to ensuring an image is discoverable. Unlike captions, which provide descriptive context, keywords act as umbrella like search terms that make it possible for buyers to find the image in a database. The challenge lies in selecting the right balance of broad and specific terms to maximize visibility. AI-powered keyword generation tools analyze an image using computer vision and cross-reference it with industry search trends, attempting to ensure that the image is tagged with optimized, relevant keywords that increase its chances of appearing in search results.
For example, if I upload my Geoffroy’s tamarin image to an AI-powered stock photography tool without any input or instructions, it might generate the following keywords:
• Basic Species Keywords: Geoffroy’s tamarin, Saguinus geoffroyi, New World monkey, small primate.
• Habitat Keywords: Lowland rainforest, tropical jungle, canopy-dwelling species, Central American wildlife.
• Behavior Keywords: Arboreal, climbing, feeding, foraging, group-living primate.
• Conservation Keywords: Endangered species, biodiversity, wildlife conservation, deforestation impact.
By using AI-driven keywording, photographers no longer must manually brainstorm search terms for each image. Instead, AI works to ensure maximum discoverability by suggesting terms that align with current buyer search behavior, industry metadata standards, and SEO best practices.
Stock photography platforms typically use one of two types of keyword strategies: controlled vocabulary or long-tail keywords.
1. Controlled Vocabulary: Agencies like Getty Images rely on pre-approved keyword lists to maintain consistency across their massive photo libraries. AI-powered tools can ensure that keywords match the agency’s structured taxonomy, preventing variations that could impact search results. For example, an image of a Geoffroy’s tamarin would always be tagged with “New World monkey” instead of “small rainforest primate.”
2. Long-Tail Keywords: Agencies like Shutterstock and Adobe Stock allow for longer, more descriptive keyword phrases that reflect how buyers actually search for images. AI analyzes search behavior trends and generates detailed keyword phrases, such as:
• Endangered Panamanian primate in tropical rainforest
• Arboreal tamarin feeding in jungle canopy
• New World monkey species native to Central America
I use Getty Images to sell some of my stock photography. But the controlled vocabulary (CV) used by Getty Images is a giant pain in the ass to work with. Their list of accepted keywords is not public, and it’s constantly changing. Try as I might to come up with 30-50 solid keywords that I believe Getty will accept, including the name of the animal, more often than not only 5-10 match their CV. And when this happens, it can take some time to work through this problem.
Case in point, I recently had to send a support ticket asking Getty Images to add Northern Pygmy Owl to their controlled vocabulary. Northern Pygmy Owl! They added it a week later, but as you can see this can be extraordinarily challenging to work with.
Some AI metadata tools are designed specifically for Getty Images controlled vocabulary, however. Pixify is the software I am currently using for keywording for Getty, and I need only select the option to use Getty Images CV to generate the keywords that will be compliant. It’s not perfect by any stretch of the imagination and once generated keywords for an orange-colored eyelash viper that included citrus fruit, fall colors, and Donald Trump. But it gets you close most of the time - with proper instructions. I find that Pixify usually manages to generate between 30-40 Getty accepted keywords than could arcuately describe the photo and concepts associated with the species in the photograph – such as wildlife conservation or biodiversity.
While AI-powered keywording is undoubtedly faster and more scalable than the alternative,, human oversight remains necessary to ensure keywords are accurate, contextually relevant, and free from biases. While AI may correctly identify a Geoffroy’s tamarin as a “New World monkey,” it might also incorrectly tag it as a “common squirrel monkey” if the dataset lacks enough training examples of tamarins. Likewise, despite what your personal views may be about the current US President, keywording an eyelash viper with his name is scientifically incorrect and would only hurt the image in search rankings because of its irrelevancy to photo buyers searching for terms like Donald Trump.
New Species of Amphibian from Panama
Caption
Recently discovered in the rainforests of Panama, Dermophis donaldtrumpi is a species of caecilian—a legless, burrowing amphibian—that has captured scientific and public attention as much for its biology as for its name. Caecilians belong to the order Gymnophiona and are among the most elusive amphibians on Earth, spending most of their lives underground and rarely seen by humans. Dermophis donaldtrumpi was named in 2018 by the sustainable investment company EnviroBuild, which purchased naming rights through a charity auction to raise awareness about rainforest biodiversity. The naming was intended as a satirical commentary on former U.S. President Donald Trump’s environmental policies, specifically his stance on climate change. The species’ subterranean lifestyle and limited use of sight were cited symbolically in reference to “burying one’s head in the sand,” alluding to perceived climate denial. This species of caecilians also survives by allowing its offspring to eat from its dead skin, which was suggested to be a form of nepotism. Although the name sparked controversy, it also brought unexpected visibility to the hidden world of caecilians and the urgent need to protect their rapidly disappearing habitats. As Panama’s forests face growing threats from deforestation and climate instability, the story of Dermophis donaldtrumpi serves as both a scientific milestone and a cautionary environmental message.
Keywords
amphibian, animal wildlife, animals in the wild, biodiversity, ecosystem, endemic - species, evolution, nature, tropical climate, tropical rainforest, vertebrate, wildlife observation, close-up, outdoors, forest, herpetology, animal skin, one animal, animal themes, forest floor, wildlife conservation, climate change, Donald Trump, Bocas del Toro, environmental conservation, Caribbean, Panama, blind snake, discovery
New Species of Amphibian from Panama
Caption
Recently discovered in the rainforests of Panama, Dermophis donaldtrumpi is a species of caecilian—a legless, burrowing amphibian—that has captured scientific and public attention as much for its biology as for its name. Caecilians belong to the order Gymnophiona and are among the most elusive amphibians on Earth, spending most of their lives underground and rarely seen by humans. Dermophis donaldtrumpi was named in 2018 by the sustainable investment company EnviroBuild, which purchased naming rights through a charity auction to raise awareness about rainforest biodiversity. The naming was intended as a satirical commentary on former U.S. President Donald Trump’s environmental policies, specifically his stance on climate change. The species’ subterranean lifestyle and limited use of sight were cited symbolically in reference to “burying one’s head in the sand,” alluding to perceived climate denial. This species of caecilians also survives by allowing its offspring to eat from its dead skin, which was suggested to be a form of nepotism. Although the name sparked controversy, it also brought unexpected visibility to the hidden world of caecilians and the urgent need to protect their rapidly disappearing habitats. As Panama’s forests face growing threats from deforestation and climate instability, the story of Dermophis donaldtrumpi serves as both a scientific milestone and a cautionary environmental message.
Keywords
amphibian, animal wildlife, animals in the wild, biodiversity, ecosystem, endemic - species, evolution, nature, tropical climate, tropical rainforest, vertebrate, wildlife observation, close-up, outdoors, forest, herpetology, animal skin, one animal, animal themes, forest floor, wildlife conservation, climate change, Donald Trump, Bocas del Toro, environmental conservation, Caribbean, Panama, blind snake, discovery
Case Studies
AI-Powered Tagging for Conservation - iNaturalist and the Geoffroy’s Tamarin
iNaturalist, a citizen science platform co-run by the California Academy of Sciences and the National Geographic Society, has been at the forefront of using AI for species identification and metadata tagging for years. Photographers, researchers, and wildlife enthusiasts upload images of plants and animals, which are then automatically analyzed by an AI-powered computer vision system trained on millions of previously identified images. This information is then used by researchers worldwide.
A few years ago, I decided to test iNaturalist’s AI-driven metadata capabilities with my own photograph of a Geoffroy’s tamarin from Panama. After uploading a series of images to the platform, I watched as the system quickly analyzed them and suggested the species as Saguinus geoffroyi, which is the correct scientific name for the tamarin. The program also generated location-based metadata, recognizing the images as being taken in a lowland tropical rainforest habitat in Central America. I was then able to add a little additional information about the exact location where I photographed the tamarin.
This AI-powered tagging isn’t just a convenience for photographers; as it’s playing an essential role in global conservation efforts, helping researchers monitor species populations and their shifting distributions due to deforestation and climate change through meta-analyses of the data compiled through the app. As someone who regularly pours over the scientific literature on species and places while researching for articles I write, I am always excited to see how citizen science (or community science) apps like this are being used by researchers, allowing wildlife and nature photographers around the world to make real and tangible contributions to academic research and conservation.
Case Study 2: AI-Driven Image Organization in National Geographic’s Archive
For decades, National Geographic has been one of the most influential publishers of wildlife photography in the world. With a vast archive of millions of images, organizing and retrieving content has always been a challenge. Imagine being an editor at National Geographic, searching for images of Geoffroy’s tamarins for an upcoming article about protecting the watershed that feeds the Panama Canal. In the past, this would require manually searching through archives, relying on photographers and photo editors before you to have inputted the right keywords and captions, and hoping the correct images surfaced. If an image wasn’t tagged with Geoffroy’s tamarin but instead with something generic like small monkey, it might never be found. Likewise, if the photo was keyworded with tropical rainforest instead of Soberanía or Chagres National Park, the photo would likely never make it out of the archives.
To solve this problem, National Geographic integrated AI-powered image recognition tools that automatically scan archived images, identify subjects, and generate new metadata based on both visual and contextual cues. This means that even older, poorly tagged images are now being properly categorized, making them searchable. AI now scans archived images for the Geographic that detects species, locations, and even behaviors like feeding, climbing, or social interactions. The system generates accurate and consistent metadata, ensuring that all relevant images appear in searches allowing editors and researchers to find images faster, reducing the time spent on manual database searches.
National Geographic’s use of AI for analyzing images and generating metadata is exactly the way in which working photographers are beginning to utilize similar programs for their own needs. I currently have over 10,000 images of wildlife just from Panama that survived the culling process but have languish away in my archives simply because I didn’t have the time to get around to spending countless hours identifying obscure species or simply fleshing out the metadata to make those files commercially sellable. But with the use of programs like Pixify for keywording and ChatGTP to create captions and titles in the style and format that I want, these images are beginning to resurface. And I now find myself slowly chipping away at that library of photos that might as well have been deleted from my archive otherwise.
Strawberry Poison Frog in Cup Mushroom on Panamanian Rainforest
Caption
Perched delicately inside the basin of a cup mushroom on the damp forest floor of Bastimentos Island, this striking red strawberry poison dart frog (Oophaga pumilio) exemplifies the extraordinary biodiversity of Panama’s Bocas del Toro archipelago. Local populations of O. pumilio across the islands exhibit an astonishing array of color morphs, ranging from tangerine orange and electric blue to leopard print and nearly every conceivable combination of colors and spots imaginable. This phenomenon, known as allopatric polymorphism, is a textbook example of how geographic isolation and microhabitat variation can drive rapid evolutionary diversification within a single species. Bastimentos Island, protected as part of the Isla Bastimentos National Marine Park, is one of the hotspots for this evolutionary process, where rainforest fragmentation, elevation gradients, and genetic drift have contributed to the emergence of highly localized morphs. The frogs’ toxicity—derived from their diet of alkaloid-rich ants and mites—makes their brilliant colors a warning to predators, a strategy known as aposematism.
Keywords
amphibian, animal wildlife, animals in the wild, biodiversity, bioreserve,dense forest, ecosystem, endemic - species, environment, forest floor, frog, natural condition, rainforest, small, threatened species, tropical rainforest, wilderness, wildlife observation, fungus, herpetology, mushroom, poison arrow frog, poisonous, symbiotic relationship, no people, nature, outdoors, close-up, macrophotography, tropical climate, strawberry poison dart frog, Bocas del Toro, Island Bastimentos National Marine Park,
The Pandora’s Box: AI Scraping and Copyright Concerns
As AI-driven metadata tools become more widely used in wildlife photography, one of the most pressing ethical concerns is the issue of AI scraping images—the unauthorized use of photographers’ work to train artificial intelligence models. Many AI systems require massive datasets to improve their accuracy in identifying species, generating captions, and suggesting keywords. To build these datasets, AI developers often pull images from publicly accessible sources across the internet. However, this raises serious copyright and intellectual property concerns, particularly for photographers whose images are being used without their knowledge or consent. Some AI developers argue that scraping public images falls under “fair use”—a legal concept that allows for the limited use of copyrighted materials for purposes like research or education. However, when AI-generated metadata tools are later commercialized and used by stock agencies, research institutions, or corporations, photographers may find themselves unintentionally contributing to a system that competes with their own work.
Unfortunately, at present, there are few mechanisms in place to allow photographers to opt out of having their images scraped by AI systems. Unlike written content, where authors can use robots.txt files to prevent search engines from indexing their work, there is no universal framework for preventing AI models from using publicly available images for training purposes.
Some major tech companies have started introducing opt-out options for artists and photographers, but enforcement remains inconsistent. Additionally, legal battles over AI-generated content are still in their early stages, and the courts have yet to establish clear precedents on whether AI scraping constitutes copyright infringement or falls under fair use.
However, there are steps that photographers can take to protect their work:
1. Using Watermarks and Image Embedding Protections – While watermarks do not prevent scraping entirely, they can deter unauthorized use of images and provide a visible claim to ownership.
2. Uploading Low-Resolution Versions of Images – Some photographers choose to share only compressed versions of their work online, reducing the quality of any scraped images used for AI training.
3. Monitoring Image Use with Reverse Image Search Tools – Services like Google Reverse Image Search or TinEye can help photographers track where their images are appearing on the web.
4. Advocating for AI Ethics and Copyright Legislation – As AI continues to evolve, photographers and legal experts need to push for stronger copyright protections that prevent AI models from being trained on copyrighted images without permission.
AI scraping and copyright concerns are valid and complex issues that need to be addressed as AI continues to shape the future of wildlife photography. The goal should not be to reject AI entirely, but rather to ensure that it is used ethically, transparently, and in a way that supports photographers rather than diminishing their work.
For me, using programs like Pixify and ChatGTP, with appropriate instructions, has transformed my workflow and dramatically increased sales of images online. Stock photography, for instance, has been and always will be a numbers game. The more high-quality photos you have in the stock photography market, the more money you will make on a monthly or annual basis. If using metadata generation tools can quadruple the number of stock images a photographer has represented by an agency, then it stands to reason that income from that agency should also quadruple.
But even when we display images on our own websites, as I mentioned before, search engines like Google are “crawling” the metadata that is embedded within the images we put online. Editors and photo buyers routinely contact me regarding licensing images that popped up with simple Google searches. Titles, keywords, and those oh-so-important captions should be considered SEO on our website, even if they are only buried in the metadata of a photo, that will both help drive traffic as well as potential sales.
AI-driven metadata has the potential to streamline workflows, improve image discoverability, and contribute to conservation efforts, but it must be implemented responsibly. If AI developers, stock agencies, and legal frameworks can adapt to protect photographers’ rights, then AI can serve as a powerful tool rather than a disruptive force, helping photographers reach broader audiences while maintaining control over their own work.
Built with Shorthand