For many birdwatchers, documenting species outside their normal range is a challenge. In the UK, this discovery often makes national headlines. For example, the Western Great Heron, commonly found in Africa and southern Europe, was spotted in a seaside town in north Wales in June and was widely praised on birding forums.
But a new scourge of AI slop threatens to destroy the fun.
Scientists are appealing to birdwatchers to limit the use of AI when editing images, fearing it could undermine the credibility of popular citizen science platforms such as iNaturalist and Macaulay Library, which are routinely used in scientific research to monitor species ranges.
The rise of generative AI platforms like ChatGPT and Google Gemini has led to a proliferation of fake and enhanced images of rare and famous birds across wildlife photography forums. Users can create high-quality fake images in seconds, or enhance photos by asking the AI to remove branches or leaves that may obscure parts of the subject. This can result in unintentionally breaking changes.
In a recent commentary in the scientific journal Nature, researchers warned that hundreds of fake images have already been discovered in common databases for recording species. They said the true scale of the problem is unknown and much may go undetected, potentially contaminating records collected by the public.
“What I realized when I looked at Facebook recently is that a huge amount of wildlife photos are now simply AI-generated images,” said Dr. Alexander Lees, an ecologist at Manchester Metropolitan University who wrote the journal article. “The idea that we might be able to use these pictures to help us understand where species exist in space and time is very challenging.”
Outright hoaxes remain rare and often easy to spot, Leeds said. Although no one would be fooled by a sighting of a toucan in Siberia, he said birdwatchers often use AI to edit and improve images, and algorithms can introduce parts of different bird species to create new images.
Leeds points to erroneous sightings of red-winged blackbirds in central Brazil. This bird is usually found in North America and had never been seen in this part of Brazil until the sighting was reported. The bird was actually an epaulette oriole, a common bird species in the New World, but the photographer asked the AI platform to make the photo “look better” and a red-winged spider was added, resulting in a false sighting.
“Wildlife photographers can be very obsessive about taking beautiful photos, but using AI to edit images runs the risk of potentially causing real problems down the road,” Leeds said.
Citizen science groups are still working to understand the scale of the problem. On iNaturalist, a social network where nature lovers can record sightings of plants and animals, only 1,400 of the more than 610 million images on the platform have been flagged for AI use. Dozens of discoveries are being made using citizen science, from monitoring how plants and animals move in response to climate change to documenting new behaviors in species.
Tony Iwane, co-author of the Nature paper and community support director at iNaturalist, said most of these cases were unlikely to be malicious, but urged users to be cautious.
“On platforms like ours, ordinary people are posting information that scientists probably don’t have access to at scale. It’s also like a sensor that tells us what’s happening on the planet in real time. Are plants blooming earlier? Are species moving north as the climate warms? “The more we know about a species’ habitat, the more informed we are as conservationists. But the information needs to be accurate,” he said.
