Balancing AI Photography and Conservation in Diving by Alex Brylske, Ph.D

Over the past year, however, a new complication has emerged, one that fundamentally alters the stakes of our visual storytelling. AI-generated wildlife images and videos are now so realistic, common, and easy to produce that they are changing how the public perceives animals, ecosystems, and risk. This has major implications not only for land conservation but also for the oceans and for everyone involved in diving, tourism, and marine protection.

The question is no longer just: What kind of pictures should we take? It is now also: How do we maintain trust when pictures themselves can no longer be trusted?

This issue was recently highlighted by Rhett Butler, founder and CEO of the online magazine Mongabay, and Mongabay contributor Sharon Guynup. Their views have significant implications for the ocean and for everyone who works at the intersection of diving, tourism, and marine conservation.

Butler’s article highlights the key issue: “AI-generated wildlife photos make conservation harder. Wildlife conservation relies on a shared understanding of reality, but as artificial images blur that line, conservation efforts become more challenging.” He points out that wildlife imagery has always been exaggerated: staged shark feeds, carefully composed reefs, and ‘once-in-a-lifetime’ encounters that somehow happen every week in marketing brochures. What has changed is the ‘speed, scale, plausibility, and ease’ with which AI now enables anyone  often with no real connection to wildlife, to create convincing scenes. To experts, the errors may be obvious. To the public, they are not.

In her article, Guynup explains how this unfolds in practice: “Fake footage distorts public understanding of animal behavior, making dangerous encounters seem normal or portraying wildlife as greater threats than they really are.” She describes AI-generated big-cat attacks, fake leopard sightings in Indian cities, and even false CCTV-style tiger killings. Authorities have wasted valuable resources investigating scenes that never happened, leaving local communities more frightened and less trusting.

Both Butler and Guynup also highlight a second, more subtle distortion: cute, anthropomorphized wildlife  tigers lounging on human backs, ‘pet-like’ animals cuddling with people, or exotic species decoratively arranged in living rooms. These images normalize close contact with and ownership of animals that are neither safe nor ethical, and they fuel the exotic pet trade.

None of this is limited to terrestrial environments. It’s easy to imagine (and, in some parts of the internet, already see) AI-generated underwater scenes: whale sharks in areas where they don’t usually live, orcas playfully nudging toddlers, manta rays like household pets in pools, or ridiculously perfect coral reefs promoted for locations whose ecosystems collapsed years ago.

For dive professionals, the risk is twofold: AI can exaggerate dangers when they are minimal and create false perceptions of health and beauty as they fade quickly.

In addition to attention from popular media, a recent article in the peer-reviewed scientific journal Conservation Biology further explores the issue. Professor Jose Guerrero-Casado from Spain’s University of Cordoba and his colleagues explain why AI wildlife fakes spread so effectively, and why their consequences can be especially harmful to conservation. The authors list three traits of modern society that worsen the problem.

First, the widespread influence of social media and realistic portrayals, whether true or fake, shapes attitudes toward wildlife, and engagement incentives (‘more clicks’) encourage sharing the most shareable images rather than the most truthful. Second, while promoting empathy for animals is positive, AI-driven narratives often exaggerate ‘human-like’ emotions and relationships, leading to widespread misconceptions about animal behavior and ecology. Third, and perhaps most relevant to our industry, a growing disconnect from nature means fewer people have enough direct experience to judge whether a scene is plausible, making them more susceptible to being misled.

Under these conditions, misinformation is not just ‘wrong information.’ It becomes a substitute reality, one that can be produced at scale, tailored to audience preferences, and spread faster than corrections.

In “Why Take Bad Pictures?” I argued that our industry’s obsession with perfect underwater images can inadvertently mislead: “In striving to create a beautiful image, one can convey the false impression that the aquatic world is doing just fine… breathtaking images often do not reflect the reality of an often-degraded ecosystem just outside the viewfinder.”

My proposal was straightforward: celebrate beauty, but purposefully include ‘bad pictures,’ images of coral bleaching, disease, ghost nets, plastic pollution, dynamite damage, and overfished reefs. These images make our storytelling more honest and more impactful for conservation.

However, the AI era demands a further step. It’s no longer enough for our images to be accurate; they must also be trusted as real. When AI can generate both ‘bad’ and ‘good’ scenes from a text prompt, authentic conservation photography, and the dive professionals who create it, becomes even more valuable. But so does our responsibility.

As Butler warns, the spread of fake imagery imposes institutional costs: “Government agencies and conservation groups are forced to divert time and resources to debunking viral content… As manipulated imagery becomes more common, genuine evidence, from camera traps to field photographs and documented encounters, may be met with skepticism.”

Guerrero-Casado and colleagues emphasize an important issue: AI-generated wildlife content can also distort public perceptions of abundance and vulnerability. If digital platforms are filled with realistic yet fake portrayals, people might believe that threatened species are faring better than they actually are, reducing the sense of urgency, donations, and volunteer efforts. In essence, fakes not only create false beliefs but can also drain motivation and weaken real-world conservation efforts.

Practical Guidelines for Dive Professionals: For those of us managing dive centers, liveaboards, training agencies, and media outlets, these issues are not just theoretical. They influence how we promote destinations, educate guests, support citizen science, and build, or weaken, trust. Here are some ways we can address these concerns:

1. Create and promote an ‘authenticity code’ (similar to the UK’s Mammal Society’s rule to reject any AI-generated nature images). The code should specify:

• Ensure that images used in training materials, briefings, and conservation campaigns are authentic and not AI-generated composites.
• If AI is used for illustration, such as conceptual graphics, clearly indicate it as AI-generated. This supports the increasing focus on identifying AI-created content.
• Avoid using AI to artificially enhance underwater wildlife behavior that never actually occurred or to manipulate the context, such as moving animals closer to humans, changing species mix, or ‘cleaning up’ background damage. This is not anti-technology. The guiding principle should be: When it comes to wildlife imagery, AI is for analysis, not for fabrication.

2. Make ‘Bad Pictures’ a systematic part of your visual curriculum, with verification in mind. Build a conservation photography library documenting:

• Coral bleaching and disease
• Physical damage (anchors, fins, storms)
• Marine debris and ghost gear
• Overfishing and loss of apex predators
• Coastal development impacts and sedimentation.

But in an era of skepticism, we must also build a verification library:

• Keep original RAW files and metadata; be ready to present unedited versions in professional or policy settings.
• When contributing to citizen science platforms, carefully follow verification and metadata protocols. As Dr. Guerrero-Casado warns, distorted or incorrect depictions can undermine species identification; that risk is amplified when the public, and even datasets, are flooded with plausible fakes.
• During briefings and post-dive talks, explain why you’re showing less-than-pretty images. Link them clearly to climate change, overfishing, and pollution, and to actions guests can take.

3. Teach ‘Media Literacy’ (especially to the audience disconnected from nature). Emphasize education as an urgent response, particularly since consistent global regulation is unlikely in the near future. Dive pros are uniquely qualified to teach this because we can combine images with firsthand interpretations.

In advanced and professional-level programs, include brief modules on:

• How wildlife is misrepresented online, through both attack videos and ‘cuddly’ content.
• Simple plausibility checks: Is this behavior plausible? Is this species found here? Does this interaction seem safe or normal?
• Why anthropomorphic storytelling is appealing, and where it clashes with biology.

This isn’t about turning divers into cynics; it’s about equipping them to become knowledgeable, discerning ambassadors.

4. Guard against distortions. The Mongabay articles highlight two opposite distortions from AI imagery: 1) Fear-inflating content (fake attacks that provoke hostility and persecution); and 2) risk-minimizing content (cute ‘friendship’ content that normalizes unsafe contact and fuels exotic pet demand). Fear-inducing content might be fabricated shark attacks or AI clips of ‘rogue’ groupers attacking divers. Risk-minimizing content might show dolphins behaving like domesticated pets or reef sharks being hugged by children.

A third distortion comes from Dr. Guerrero-Casado’s article: false location claims, posting images and asserting they were taken where the species doesn’t exist. This can mislead the public about native biodiversity and even increase tourist pressure on fragile sites. In an industry already driven by ‘bucket list’ travel, a viral fake can lead real people into authentic habitats, causing genuine ecological problems.

5. Protect the data stream, not just the story. We often see images as tools for marketing or education. Yet, conservation science increasingly depends on online content to answer ecological questions, such as occurrence, behavior, and distribution. AI-generated imagery can pollute this ‘digital ecology’ by making the filtering process more complex. For dive professionals who post frequently, this means a responsibility not only to audiences but also to future data users: be precise about the date, location, and species identification; avoid vague captions; and keep source files when possible.

6. Be aware of the ‘charisma bias.’ The Guerrero-Casado article suggests that AI wildlife content will tend to focus on mammals because they are most engaging on social media, which can distort public attention and conservation funding. Marine ecosystems have their own charisma hierarchy, sharks, turtles, dolphins, whales, while less ‘engaging’ taxa (invertebrates, seagrass communities, many reef fishes) are already ignored. One way to address this is to intentionally diversify what we photograph, teach, and celebrate: not only apex predators but also habitats and overlooked organisms that genuinely help reefs function.

Marketing in a Post-Truth Marketplace: Even before AI, our industry already tended to cherry-pick the truth: brochure-perfect reefs, endless sunny days, and wildlife schedules that match theme park timetables. Now, AI encourages us to take those tendencies even further. But in a world flooded with incredible images, credibility remains important. For operators and agencies, that might mean:

• Label your visual content as authentic field photography, with behind-the-scenes notes on how and when it was captured.
• Occasionally show ‘shoulder shots’ of damaged environments, not to deter customers, but to attract divers who want to understand and help.
• Train staff in visual ethics and transparent disclosure to establish your operation as a trusted source instead of merely another marketing image or social media post.

In my earlier “Bad Pictures” article, I concluded: “Photography can be a tool to either conceal reality or inspire the action needed to return the ocean to a healthier state. Don’t just take pictures; take pictures with a purpose.” Butler and Guynup extend that imperative. Guerrero-Casado and colleagues broaden it further: the problem is not merely fake images, but how they exploit social media dynamics, anthropomorphism, and disconnection from nature, distorting knowledge, attention, tourism patterns, and even the data streams conservation relies on.

Paid Advertisement

For our dive industry, the response cannot be to deceive or to retreat from imagery. It is to embrace our role as frontline witnesses: to document honestly, label transparently, teach media literacy, and market responsibly.

We are entering a time when the question ‘Is this real?’ will be asked about every compelling marine image we publish. If we do our jobs well now, by using authenticity cues, including ‘bad pictures’ in conservation storytelling, and educating our audiences about AI, then our answer can be: Yes, and here’s why it matters. In a world of synthetic oceans, the real sea, and genuine photographs of it, need more professional advocates than ever before.

Point/CounterPoint by Marty Snyderman

A few months ago, Alex Brylske asked my opinion about how AI is used in photography and if I would be willing to share my thoughts about the accompanying article. I knew then that I was of several minds, but because Alex is Alex, I promised him I would give it a go. So, over the past few months I have tried to solidify my thoughts.

That said, I remain conflicted. Alex’s piece specifically deals with how AI is used in the field of conservation. I understand that along with Alex, some conservationists want all photographers and publishers to clearly and honestly say whether AI was used to make or enhance their images so that pictures from the field are accurate and real. I think the intention is noble. In a perfect idealistic world that might happen, but my opinion is that for a variety of reasons a significant percentage of image makers won’t comply. Some won’t be aware – at least not yet, some won’t care, some won’t want to take the time, some will fail to label their images in a way that shows if AI was used, as time passes some people will forget whether AI was used or not, etc. In some respects, it is analogous to having laws regarding safe driving and therefore assume, no one will speed, roll through stop signs, or drive under the influence, and everyone will always use their seatbelt.

Another factor is the issue of “where do you draw the line?” This question is not new. Some people seem to think that if you make a photograph using film that it is pure. However, the type of film used greatly influences the colors in a photograph. My bet is that photographers have used the film they feel will make their images look the way they want them to, so their choice of film might saturate colors or enhance certain colors at the expense of others. The use of Ektachrome films was often popular to make less appealing green water take on a bluer appearance. People tended to praise, not criticize, these photographers for the way they went about their work.

I don’t think this will come as a great surprise, but it is easy to overlook: When making films, it is common practice for the most highly acclaimed natural history production companies to add sound effects that were not present at the time a clip or scene was made. The same is obviously true for music. Do we ask for reality or entertainment? And some scenes in “blue chip” underwater films are made with captive animals, the use of bait, or footage that was acquired at another time and location to complete a sequence.

From a business perspective when advertising, entities “put their best foot forward.” Given a choice do you use images of blooming flowers of those that have lost their petals. Do you expect dive resorts and boats to share images with the vibrant reefs or those that have been reduced to rubble by bleaching, pollution, stony coral tissue disease etc? All might be honest, but not all will be used. That’s business and we would be foolish to expect otherwise.

AI is here to stay. My opinion is: in the world we live in some people will use AI in beautifully artistic ways with absolutely no constraints. That’s artistic license. Some will use AI in their images in ways they believe to be ethical, but people’s feelings about ethics differ greatly. Others won’t care or feel like an unnecessary burden is being placed on them.

Some profess that the best path forward is to create an accepted symbol that designates that AI was not used when making an image. But would that mean a photographer could not use Photoshop, Lightroom, or other editing software to produce those images, or only certain tools? If so, which ones? For me, the answers are not black and white. I feel certain there are those who will strongly disagree with my thoughts. They are entitled to. I like to think I am open–minded and my opinions might evolve over time.