AI-Edited Bird Photos Threaten Scientific Research, Scientists Warn
Scientists have warned that AI-enhanced wildlife photographs, increasingly shared on citizen science platforms like iNaturalist and the Macaulay Library, could undermine the reliability of bird research. A commentary in the journal Nature, cited by The Guardian, reports that hundreds of fake images have already been detected on species-recording databases, though the full scale remains unknown. Dr. Alexander Lees, an ecologist at Manchester Metropolitan University and author of the article, highlighted a case where a photographer used AI to improve an image of an epaulet oriole in central Brazil, inadvertently adding features of a red-winged blackbird—a species never recorded there—creating a false record. Lees noted that many wildlife photographers are "obsessed with getting a beautiful photo" but risk altering crucial details. He observed that a huge volume of wildlife photos on Facebook are now AI-generated, complicating efforts to track species distributions. On iNaturalist, only 1,400 of over 610 million images have been flagged for AI use, indicating the challenge of detection.
Key facts
- Scientists warn AI-edited bird photos threaten research reliability.
- Commentary published in Nature journal.
- Hundreds of fake images found on citizen science databases.
- Full extent of AI-altered images remains unclear.
- Dr. Alexander Lees of Manchester Metropolitan University authored the commentary.
- False sighting of red-winged blackbird in central Brazil due to AI editing.
- Red-winged blackbird is typically found in North America.
- Only 1,400 of 610 million iNaturalist images flagged for AI use.
Entities
Institutions
- The Guardian
- iNaturalist
- Macaulay Library
- Nature
- Manchester Metropolitan University
- PetaPixel
- Depositphotos
- KEH.com
- Moment
- FJ Westcott
- Blue Moon Camera and Machine
Locations
- central Brazil
- North America