Nikon has disqualified the original first-place video in its Small World in Motion microscopy contest after concluding that the entry violated its rules on generative AI. The company promoted the remaining entries and named a video by Vietnam’s Nguyen Nam Nhat as the new winner.
The removed work, submitted by Ning Xu of Tsinghua University, depicted cilia moving in the airway of a child with a rare respiratory disorder. Researchers raised concerns about structures appearing and disappearing in ways they said did not match cellular biology, as well as markers in source files that resembled AI watermarks.
Xu said an unsupervised neural-network method helped distinguish and visualize features in reconstructed grayscale images. He denied using AI to generate the experimental movie, the cilia or their motion. Nikon’s statement does not accuse him of fabricating the underlying experiment; it says only that the submitted video did not meet the competition’s eligibility rules.
The case exposes a difficult boundary in scientific imaging, where machine learning may be used for reconstruction, denoising or visualization but can also introduce details that look like observations. Nikon says it will revisit its rules and evaluation procedures. Clear disclosure of processing steps and preservation of raw data will be essential when judges cannot infer provenance from the final image alone.