Source: Popular Science

Researchers have discovered that pigeons can accurately detect subtle signs of lung disease in CT scans, offering insights that could improve artificial intelligence tools for medical imaging. In a recent study, six pigeons were trained to distinguish scans with lung nodules from normal scans, achieving impressive accuracy and successfully identifying abnormalities in previously unseen images. Remarkably, the birds also recognized emphysema and ground-glass nodules without specific training, suggesting they detect visual patterns that humans may unconsciously perceive. 
The research builds on earlier findings showing that radiologists’ eyes and brains often notice suspicious abnormalities even when they consciously classify scans as normal. Scientists hope to use eye tracking and physiological responses, such as pupil dilation, to train AI systems that complement radiologists by identifying overlooked abnormalities. Rather than replacing physicians, the technology could enhance diagnostic accuracy and potentially be adapted for other fields, including cardiology, airport security screening, and art authentication.