Artificial intelligence is increasingly complicating efforts to document rare bird sightings, prompting researchers to warn that manipulated photographs could weaken the scientific value of major citizen-science databases, The Guardian reported.
Platforms such as iNaturalist and the Macaulay Library allow members of the public to upload photographs and location data for birds, plants, and other wildlife. Researchers regularly rely on these records to track changes in species distribution, habitat use, migration, and behavior.
However, scientists say the rapid spread of AI-generated and AI-enhanced wildlife images is making it harder to distinguish genuine observations from altered material.
The concern is not limited to entirely fabricated photographs. Even relatively minor edits, such as removing leaves, branches, or other objects obscuring an animal, can lead generative AI systems to invent details that were not present in the original image.
Rare sightings particularly vulnerable
Unusual bird sightings often attract significant attention among birdwatchers, particularly when a species appears far outside its established geographic range.
In June, the appearance of a western reef heron in a coastal town in north Wales generated widespread interest. The species is normally associated with Africa and parts of southern Europe, making the sighting especially notable among British birding communities.
Researchers now fear that the growing volume of artificial or heavily modified photographs could cast doubt on such records, particularly when the sighting involves a species that has never previously been documented in a certain region.
A commentary published in the journal Nature warned that hundreds of false images have already been identified across widely used wildlife-recording platforms. The researchers said the actual number may be significantly higher because sophisticated alterations can escape detection.
“My experience of looking at Facebook these days is that a huge volume of wildlife photos now are simply AI-generated imagery,” said Dr Alexander Lees, an ecologist at Manchester Metropolitan University who authored the journal article.
The idea that we could maybe use those photos to help us understand where species are in space and time is very difficult.
Enhancement tools can change species
According to Lees, deliberately staged hoaxes remain uncommon and are generally easy to identify when they involve implausible sightings. More subtle alterations, however, present a greater threat.
Bird photographers may use AI tools to sharpen images, improve lighting, remove visual obstructions, or make a subject appear clearer. During that process, the software can introduce markings, colors, or anatomical features associated with an entirely different bird.
One case involved a photograph taken in central Brazil that was initially presented as a red-winged blackbird, a species native primarily to North America and not previously recorded in that region of Brazil.
The bird was later identified as an epaulet oriole, a species commonly found in the Americas. The photographer had reportedly used an AI tool to improve the image, after which the software added characteristics resembling those of a red-winged blackbird.
“Wildlife photographers can be quite obsessed with getting a beautiful photo, but there’s a risk that the image might actually cause problems down the line when AI has been used to edit it,” said Lees.
The incident demonstrates how an apparently harmless attempt to improve an image can create a false scientific record, even when the photographer has no intention of misleading others.
Citizen science depends on trustworthy records
Citizen-science platforms have become an increasingly important source of environmental information because they collect observations from people across large geographic areas and over long periods.
These records can help researchers identify changes that would be difficult or expensive to document through conventional fieldwork alone. They have been used to examine whether species are shifting northward as temperatures rise, whether plants are flowering earlier, and whether animals are displaying previously undocumented behavior.
On iNaturalist, only around 1,400 images have so far been flagged for AI use out of more than 610 million uploaded photographs. Researchers caution, however, that the figure may not represent the true scale of the problem because many altered images may never be identified.
Tony Iwane, iNaturalist’s director of community support and a co-author of the Nature commentary, said most questionable submissions were probably not intended as deliberate deception.
“On platforms like ours, regular people are posting information that a scientist could probably never get at scale. It is also almost like a sensor of what is happening on Earth in real time: are plants flowering early? Are species moving north as the climate warms? The more we know about where species are, the better informed we can be as conservationists. But the information needs to be accurate,” he said.
Scientists are urging wildlife photographers and birdwatchers to avoid using generative AI when preparing images intended for scientific databases. They are also calling for greater transparency whenever digital editing has been used.
Without clear standards, researchers warn that AI-modified images could gradually contaminate public wildlife records, making it harder to determine whether a rare species was genuinely observed or digitally created.