AI-generated wildlife images are endangering animals and people, experts warn
Conservationists warn that hyper-realistic AI-generated wildlife images circulating on social media distort public perception of animals, fueling exploitation and unsafe human encounters.
Intelligence analysis by Llama
Fact-checkers have debunked dozens of AI-generated wildlife images, from an orangutan with leopard cubs in Sabah to a tree-climbing elephant in Myanmar. Experts say the synthetic visuals encourage risky interactions with wildlife and erode trust in real conservation footage.
Some people use computers to make fake pictures of wild animals doing cute things, like a monkey holding baby leopards. The pictures look so real that others share them and forget that real wild animals are not pets. That can trick people into getting too close to dangerous animals, and it makes it harder to know which animal photos online are real.
Analysis
Sabah
A viral black-and-white image of an orangutan cradling leopard cubs in the jungles of Malaysia's Sabah state has become the emblem of a new misinformation problem. The picture is entirely synthetic, yet it circulates as if authentic, accompanied by heartwarming captions across Facebook, TikTok, and X in multiple languages. Conservationists note that the same pattern recurs in other debunked images: an elephant climbing a tree to escape flooding in Myanmar, a pink dolphin leaping from Philippine waters. Each faked scene places a wild animal in a posture of human-like tenderness, blurring the line between documentary photography and engineered fiction.
Xinjiang snow leopard
The danger is not abstract. In January, a woman in China's northwest Xinjiang region was attacked by a rare snow leopard after approaching it for a photo, an incident that prompted China's forestry department to step up patrols and warn the public to maintain safe distances from wild animals. Zoology professor Jose Guerrero-Casado of Spain's University of Cordoba pointed directly to the cumulative effect of AI visuals that portray wild animals as cute or cuddly. When people lose a sense of where the boundary between species lies, he argued, they behave in ways that are unsafe for both themselves and the animals. Born Free researcher Chris Lewis added a second harm pathway: synthetic imagery increases demand for certain species as tourism photo props, social media subjects, or pets, a chain of exploitation that AI content quietly amplifies.
September study
A study co-authored by Guerrero-Casado and published in the journal Conservation Biology in September labels the rapid spread of AI misinformation a major threat to environmental efforts, warning that fabricated visuals could affect the ability of people to identify species and distort understandings of wildlife abundance and distribution. A separate commentary in Nature called for more robust authentication of submissions to citizen science platforms, which researchers mine for population and behavioral data, and for greater user education about the risks of manipulation. Meta and TikTok require users to label realistic AI content and say they deploy automatic detection tools, yet fact-checkers, including AFP, which has fact-checking contracts with both companies, have found multiple wildlife posts circulating without clear labels, even where captions acknowledge the images are fake. WWF's Jenny Roberts summed up the downstream cost: when real conservation footage, such as camera-trap video of a tiger with five cubs in China, faces skepticism, fundraising and public engagement both suffer.
Key points
- A fake image of an orangutan cradling leopard cubs in Malaysia's Sabah state has become emblematic of a wave of AI-generated wildlife misinformation on social media.
- Conservationists cite a January snow leopard attack in China's Xinjiang region as evidence that anthropomorphized AI images encourage unsafe human behavior toward wild animals.
- A September study in the journal Conservation Biology labels the spread of AI wildlife misinformation a major threat to environmental efforts and species identification.
- Experts including Jose Guerrero-Casado, Chris Lewis of Born Free, and Jenny Roberts of WWF warn that synthetic imagery fuels exploitation and erodes trust in real conservation footage.
- Meta and TikTok require labels on realistic AI content and pay AFP for fact-checking, but fact-checkers have found multiple wildlife posts circulating without clear labels.
Platforms including Meta and TikTok now require labels on realistic AI content and have signed fact-checking partnerships with AFP, which could raise the cost of distributing unlabeled synthetic wildlife imagery. Conservation organizations can also adapt by embracing provenance tools and educating audiences, as WWF's call for robust authentication of citizen science submissions suggests.
The September Conservation Biology study frames the spread of AI wildlife misinformation as a major threat, and experts note that detection systems are imperfect, with fact-checkers already finding unlabeled fake posts on major platforms. If trust in genuine conservation imagery continues to erode, fundraising may suffer and citizen science datasets risk contamination with fabricated records.