The ruling
Nikon has disqualified the winning entry in its 2026 Small World in Motion competition after concluding that the video broke the contest’s rules on generative AI.
“After thoroughly re-evaluating the video and supporting materials and consulting with members of our judging panel, Nikon has determined that the winning video did not comply with the competition rules regarding generative AI,” a Nikon spokesperson told PetaPixel on 9 October. “Nikon has adjusted the competition rankings to reflect this decision.”
The company drew a line around what it was deciding. “This decision is based solely on the submitted video’s eligibility under the rules outlined for this competition,” the spokesperson said. “It should not be interpreted as a judgment of the entrant’s professional reputation, scientific contributions, or intent.”
Nikon’s gallery now lists Nguyen Nam Nhat first, for a video of a roundworm and a single-celled Dileptus. The disqualified entrant does not appear.
What the video showed, and what others saw in it
The entry came from Ning Xu, an optical engineering researcher at the National University of Singapore. It was presented as showing the movement of cilia — the hair-like structures lining part of the airway — in tissue from a child with primary ciliary dyskinesia, a rare genetic disorder.
Scientists raised objections about the imagery itself. Edward Phelps, an associate professor in Florida, wrote on LinkedIn that the video had serious problems: “purple structures pop in and out of existence”, he wrote, and “green cilia also appear from nowhere and do not match the known size”. Others said the question could have been settled if the raw grayscale footage had been shared.
Xu acknowledged using AI in post-processing but denied breaking the rules, which state that AI-generated videos are not permitted. That is the whole dispute in one sentence: he treats post-processing as outside the prohibition, and Nikon’s panel did not.
Why this is harder than it looks
In ordinary photography the boundary between enhancement and generation is contested but at least visible. In microscopy it is neither. Raw microscope output is often faint, noisy, grayscale and in need of substantial computational treatment before a human can see anything at all — denoising, deconvolution, segmentation, false colour. Much of that pipeline now uses neural networks, and most of it predates anyone calling it AI.
A rule that says “AI-generated videos are not permitted” does not tell an entrant which parts of that chain are allowed. The failure mode is not fraud; it is a contest whose rules were written for a different technology.
Nikon says the rules need rewriting
The company said as much. “Advances in imaging and artificial intelligence continue to evolve rapidly, creating new challenges for organizations across many fields,” Nikon said. “This event has shown the need for Nikon Small World and Nikon Small World in Motion to revisit rules and evaluation procedures for future competition entries.”
What to watch
The useful outcome would be a rule that names the step rather than the technology: which operations are permitted, and what has to be submitted alongside the finished video. Several scientists in this case pointed at the same remedy — the raw file.
Science competitions are a small corner of the problem. Journals, conferences and funding bodies are all working out the same boundary, with the same vocabulary, and mostly without the option of adjusting the rankings afterwards.
