AI-enhanced microscopy video sparks debate over scientific accuracy in competition

On 15 September, Ning Xu of the National University of Singapore was announced the winner of the annual Small World in Motion competition, which showcases light-microscope videos. His submission visualises cilia movement in lung tissue taken from a child with primary ciliary dyskinesia, overlaying red, purple and blue structures that lack explanation in the competition’s official description. Shortly after the announcement, researchers such as Melanie White of the University of Queensland and Markus Sauer of the University of Würzburg raised concerns that the coloured elements may be AI-generated hallucinations rather than real biological features.

Edward Phelps of the University of Florida also highlighted inconsistencies, noting that the purple shapes resemble mitochondria in a biologically implausible arrangement. In response, Xu said the AI model was used solely for post-processing to differentiate and colour structures with similar morphology, without making anatomical claims about what those features represent. The controversy underscores the tension between aesthetic enhancement and data integrity in scientific imaging.