The creator says AI was used to make the footage more visually engaging. Researchers question whether the result still reflects what the microscope recorded.
Nikon’s Small World in Motion competition awarded its top prize on 15 September to a video of moving lung cilia, the hair-like structures that help clear the airways. Created by optical engineer Ning Xu of the National University of Singapore, it shows the cilia above red, purple and blue features whose biological interpretation has drawn criticism, Nature reports.
A University of Florida bioengineering researcher said the purple forms resemble mitochondria but appear outside cells and as large as cell nuclei—an arrangement he described as biologically impossible. The blue forms look like nuclei but do not behave like them, he said, while the red stain’s meaning is unclear.
Xu said the processing was intended to improve the appearance of the region beneath the cilia, without claiming that the rendered features represented particular anatomy. He said AI had not created the original experimental footage or the cilia’s motion: the grayscale data had already been reconstructed before AI was used to distinguish similarly shaped structures and color them.
A University of Queensland developmental biologist said AI might have introduced structures absent from the data. Scientific images serve as data, the biologist argued, and viewers need to trust that they reflect what was measured.
Nikon added an AI-processing disclosure to the entry on 22 September, saying an unsupervised model had helped with post-processing. The company said it was re-reviewing the submission and had received detailed documentation from Xu on the equipment, imaging methods and processing used to create the source video.