Back in July, an AI-generated video of a bunch of rabbits bouncing on a trampoline rocked the internet, with people claiming it was the first AI video they’d ever been hooked on. But new research suggests this type of media may actually be harmful to real wildlife.
Researchers at the University of Córdoba (UCO) in Spain have raised concerns about the growing circulation of AI-generated wildlife videos on social media and their potential impact on the public’s understanding of nature.
The GESBIO research group (José Guerrero, Francisco Sánchez, Antonio Carpio, Rocio Serrano, and Tamara Murillo) investigated some of the most widely shared examples, including the popular clip of a leopard entering a backyard where children are playing and being chased away by a pet cat.
The leopard video garnered more than 1 million likes and tens of thousands of shares on Instagram. Other viral clips show the hedgehog riding down a river on the back of a crocodile and battling a raccoon and a bear. It’s all created with artificial intelligence, but rendered in a way that looks real.
The researchers argue that the realism of these videos can distort viewers’ understanding of wildlife, especially children. Their analysis highlights several issues, including unrealistic depictions of animal behavior, the application of human-like traits to wild species, and the widening gap between ordinary people and the natural world.
“They reflect traits, behaviors, habitats, or relationships between species that are not real. For example, we see predators and prey playing. They show us animals whose human behavior is far removed from reality,” Guerrero says. He added that the widely shared video of the leopard and the cub “undermines the conservation of these species, as they would never be encountered in these conditions.”
Serrano warns that AI-generated wildlife clips are reinforcing this disconnect by portraying rare and endangered species as commonplace.
“Because these videos feature more endangered species, they create a false connection with nature, which is negative for conservation,” she explains. The impact can also extend to outdoor expectations, where children expect to encounter animals with “more charismatic or magical characteristics or behaviors.”
“If young children don’t go out into the countryside and find animals with more charismatic and magical characteristics and behaviors, it’s counterproductive in terms of connections,” she added.
Researchers argue that these distortions could have lasting effects, as children rely heavily on visual content when learning and social media is becoming a primary source of information. One such result is a growing interest in keeping exotic animals as pets.
“We are facing another serious problem: the demand for exotic species as pets,” Murillo said, warning that videos depicting wild animals as docile or sociable could encourage inappropriate ownership.
To counter these trends, the team recommends increasing media literacy so that viewers can better assess the trustworthiness of online content. They also suggest incorporating environmental education into school curricula. Sanchez emphasizes the importance of clarifying basic concepts. Media literacy needs to provide “the tools to question and verify information by relying on trusted sources,” while environmental education needs to ensure “children understand from an early age that there are no lions here.”
Their findings, published in the Conservation Issues section of Conservation Biology, represent an early attempt to assess how AI-generated media influences perceptions of biodiversity and highlight the need for additional research.
