In brief
The 2026 Small World In Motion contest, a long‑running microscopical photography competition, has been rocked by claims that its winning video, submitted by Tsinghua University researcher Dr Ning Xu, was produced with generative AI. While the entry, depicting cilia in a child’s airway, won a £2,250 prize, scientists raised concerns over anomalous shapes and unnatural motion in the footage. Dr Xu denies using AI to create the image, but admits employing AI tools for image reconstruction. Nikon maintains the footage complies with contest rules, though the investigation is ongoing. The controversy highlights growing tension between cutting‑edge AI techniques and the integrity of scientific visual media.
In October, Nikon’s long‑running Small World In Motion competition, for microscopic photographs and videos, felt the heat of scrutiny after Dr Ning Xu of Tsinghua University won the 2026 prize for a clip that showed hair‑like cilia moving inside a child’s airway.
While microscopes can produce bright, artificial colours through staining, the clip’s cilia appeared larger and moved in a way that some scientists say does not match normal biology.
Dr Xu issued a statement on LinkedIn, saying “AI was not used to generate the experimental movie, the cilia, or their motion.” He acknowledged that AI helped in “distinguishing and visualizing features in the reconstructed grayscale images.”
Some former judges, along with associate professor Edward Phelps from Florida University and PhD student Ian Donovan of UT Southwestern, raised concerns about the video’s authenticity, noting brief appearances of the cilia “pop in and out of existence” and possible watermarking of AI origins.
Nikon’s statement says it is “respectfully complying” and has supplied documentation to the inquiry. The company maintains that the footage follows contest rules, which explicitly forbid generative AI creation but allow AI for data processing.
The debate underscores a broader question for scientific media: how to balance the benefits of AI processing with the need for verifiable integrity in research imagery.
As Nikon announces its next steps, the scientific community watches closely, aware that the outcome could set precedent for how AI‑enhanced images are judged in research contexts.
