Nikon disqualifies winning microscopy video over generative AI

Nikon has disqualified Dr Ning Xu’s winning video from its Small World In Motion microscopy competition over generative AI use, BBC News reported on October 9. The decision follows scientists’ challenges to the entry’s authenticity and comes with a commitment to review the contest’s rules and procedures.

Nikon said the video ‘did not comply with the competition rules regarding generative AI’. That finding is the organiser’s stated reason for removing the winning entry, which had been presented as showing the movement of microscopic structures in a child’s airway.

According to the BBC report, Nikon reached its decision ‘after thoroughly re-evaluating the video and supporting materials and consulting with members of our judging panel’. It expressly said the disqualification was not a judgment of Xu’s professional reputation or scientific contributions.

How Nikon’s review developed

Small World In Motion is a competition for microscopic imagery. Xu’s short video was presented as depicting tiny, hair-like airway structures called cilia moving in a child with primary ciliary dyskinesia, or PCD. CNN described the submitted entry as purportedly showing abnormal cilia movement; the footage’s biological content remains disputed.

On September 26, Nikon said in a LinkedIn post that Xu had cooperated and supplied detailed technical documentation, according to CNN. Those materials covered the microscopy equipment, imaging methods and image-processing techniques used to produce the entry, giving the organiser information to examine alongside the video.

The BBC identified Xu with Tsinghua University in China. CNN’s earlier report, attributing the information to Nikon, said he had subsequently joined the National University of Singapore as a research fellow.

What Ning Xu said about using AI

During the earlier review, Xu disputed that AI had generated the experimental video, the cilia or their movement, CNN reported. He said AI was applied afterward to reconstructed grayscale data to distinguish and colour structures with similar shapes. That explanation described a processing step after reconstruction.

The BBC’s October 9 report also said Xu had acknowledged using AI but believed he had acted within the rules and had cooperated with Nikon’s investigation. His stated position therefore concerned both how AI was used and whether that use was permitted by the competition.

The BBC said it had approached Xu for comment on the disqualification. Its report contained no new response from him to the final decision, so his earlier explanation remains the available account of his position before Nikon announced the outcome.

Why scientists questioned the cilia footage

Scientists, including former competition judges, raised concerns on Nikon’s LinkedIn page that the video did not appear real, the BBC reported. One of the experts who challenged it was Dr Robert Hirst of the University of Leicester, who said he was ‘not surprised’ by the disqualification.

The BBC identified Hirst as lead scientist for the university’s NHS centre for PCD diagnosis. He said he had spent 20 years diagnosing the condition by examining cilia waveform and length, alongside cell size and shape. Those features formed the basis of his criticism of the entry.

Hirst told the BBC that the cells and cilia ‘look nothing like those from PCD patients’. He also said the controversy had affected patients, scientists, doctors and PCD support networks around the world. The report did not quantify those effects or establish clinical harm.

What Nikon’s procedural review leaves open

Nikon said it would examine the annual contest’s rules and procedures. The BBC report did not identify any adopted rule changes or give an implementation date. The announced review therefore leaves the details of future entry requirements and judging procedures unresolved.

The reports also do not establish a replacement winner, revised rankings, any prize-repayment decision or an appeal procedure. Nor do they establish the precise AI model, processing pipeline or extent of generated content in Xu’s entry. Nikon’s stated finding concerns compliance with its generative AI rules.