Nikon’s Small World in Motion contest was founded to celebrate what microscopes reveal: science rendered startlingly alive. But its latest result has become a test of where legitimate visualization ends and generative AI begins.

Last month, Ning Xu of Tsinghua University in China won with a video showing cilia beating in the airways of a child with a rare respiratory disorder. Soon afterward, researchers raised alarms over apparent structural abnormalities, cell features that seemed to appear and disappear, and what looked like AI watermarks in source files.¹

Xu acknowledged using AI “to distinguish and visualize features in the reconstructed grayscale images,” but denied using it “to generate the experimental movie, the cilia, or their motion.”¹ That distinction sits at the heart of the dispute: Xu’s account frames AI as an interpretive aid, while critics saw signs that the resulting video could no longer be trusted as a faithful scientific record.

After re-examining the video and supporting material with members of its judging panel, Nikon concluded that the entry breached its generative-AI rules and stripped it of the prize. The company said the decision was “based solely” on eligibility and should not be read as a judgment of Xu’s professional standing, scientific work or intent.¹

The rankings have now shifted. Nguyen Nam Nhat of Vietnam takes first place with footage of a tiny roundworm and a single-celled organism; the remaining finalists each move up one position.¹

Nikon’s larger concession may matter most. Rapid advances in imaging and AI, it said, have created fresh challenges and shown the need to revisit contest rules and evaluation procedures.¹ The disqualification settles this year’s prize, but not the harder question: how scientific-image competitions should police tools that can clarify a picture—or quietly remake it.