A commentary published in the journal Nature raises concerns that AI-generated and AI-enhanced images of birds are appearing on citizen science databases, potentially affecting the reliability of data used for ecological research.
Scale of the Issue
Researchers report that hundreds of potentially fake or edited images have been discovered on platforms such as iNaturalist and the Macaulay Library. The full extent of the problem remains unknown, as many altered images may go undetected.
On iNaturalist, which hosts over 610 million images, approximately 1,400 have been flagged for the use of artificial intelligence.
Specific Incident
An example cited in the commentary involves a purported sighting of a red-winged blackbird in central Brazil, a species native to North America that had never been recorded in that region.
Investigation revealed that the image was an AI-enhanced photograph of an epaulet oriole. The photographer had used an AI tool to improve the image, which inadvertently introduced features from the red-winged blackbird.
Expert Observations
“The idea that we could maybe use those photos to help us understand where species are in space and time is very difficult.”
— Dr. Alexander Lees, ecologist at Manchester Metropolitan University
Dr. Lees noted that a significant portion of wildlife photos on social media now appear to be AI-generated. He also remarked that photographers can be focused on obtaining a visually appealing image, but editing with AI risks introducing inaccuracies.
Tony Iwane, director of community support at iNaturalist and a co-author, stated that most flagged cases are not intentional attempts to deceive. He emphasized the importance of accurate data, noting that the platforms allow the public to provide information used by scientists and conservationists, but that the information must be reliable.
Context
Citizen science platforms such as iNaturalist and the Macaulay Library rely on public submissions to monitor species distribution, migration patterns, and behavioral responses to environmental changes. The authors of the commentary urge users to be vigilant about the use of AI tools when submitting images to ensure the integrity of scientific data.