A celebrated entry in one of the world's most prestigious scientific imaging competitions is now the subject of an unusual public dispute, after researchers raised concerns that the winning video may have been altered using artificial intelligence. The controversy centres on Nikon's annual Small World In Motion contest, an event that for 15 years has showcased the best photography and video captured through powerful microscopes.
The winning submission came from Dr Ning Xu, a researcher affiliated with Tsinghua University in China. According to the competition's description, the short video purports to show cilia — tiny, hair-like structures — beating inside the airways of a child suffering from a rare disease. Such footage, if genuine, would represent a striking visual record of microscopic biology in action.
However, almost as soon as the prize was announced, other scientists began questioning the authenticity of what the video depicted. Their objections focused not on the fact that image processing had taken place, but on the possibility that generative AI had been used to fabricate visual elements that were never actually present in the original sample.
Dr Xu did not reply to a request for comment from the BBC. Nevertheless, he has publicly denied breaking the contest's rules. Responding to a congratulatory post by Nikon on LinkedIn, he maintained that AI played only a limited role in his submission and that this role fell within what the competition permits. "AI was not used to generate the experimental movie, the cilia, or their motion," he wrote, according to the source account.
Xu did acknowledge that AI tools were employed "to distinguish and visualize features in the reconstructed grayscale images." He also pointed out that the vivid, artificial-looking colours in his video are common in microscopy and explicitly allowed by the contest — typically achieved through manual staining and dyes applied to the sample itself. In other words, enhancement for visibility is normal; the question is where enhancement ends and invention begins.
That distinction sits at the heart of the complaints now circulating among researchers. Critics argue that the video contains shapes and patterns of movement that do not match normal cell structures or behaviours. Some noted that the cilia appeared far larger than they should be, given the biology being depicted. Others pointed to more fundamental problems with the footage's internal logic.
Edward Phelps, an associate professor at Florida University, commented on Nikon's social media post that certain structures "pop in and out of existence" and that the cilia "appear from nowhere." Such observations, if accurate, would be difficult to reconcile with genuine microscope footage, in which physical structures persist and move according to biological constraints. Meanwhile, Ian Donovan, a PhD student at UT Southwestern Medical Centre, said the source image appeared to carry a distinctive watermark associated with AI-generated content — a potentially damning detail if verified.
The concerns have not been limited to early-career researchers. Two former judges of the competition have also voiced doubts about whether the image was real, adding institutional weight to the criticism and increasing pressure on Nikon to explain how its vetting process works.
Nikon's response so far has been cautious. The company stated that Xu was "respectfully complying" with its investigation and had supplied additional technical documents to assist the review. It also said that, at this stage, it did not see that any rules had been violated. At the same time, Nikon has promised an update soon, leaving open the possibility that its position could change as the inquiry proceeds.
The stakes extend well beyond the prize money. Dr Patrick Hickey, who finished fifth in the 2026 Small World contest, explained that the recognition associated with the award carries far more weight than the $3,000 (£2,250) that accompanies it. For scientists who spend long hours in darkened rooms peering down microscopes, he said, the chance to show their work to the public is genuinely valuable — a sentiment that underscores why allegations of fakery are being taken so seriously by the community.
Hickey also highlighted the artistic dimension of the field, describing an interesting fusion between science and art, and noting that many images produced by researchers possess abstract, almost painterly qualities. That aesthetic appeal is part of what makes the competition popular — but it also makes the boundary between legitimate visual enhancement and misleading fabrication especially fraught.
In Hickey's view, fairness to other entrants is the central issue. He pointed to the contest's explicit rules, one of which clearly prohibits the use of generative AI to produce content, and another of which requires that footage be taken under a microscope. If the winning video turned out to have been unfairly altered by AI, he said, that would be unjust to everyone else who competed under those conditions.
The episode raises broader questions about how scientific competitions verify submissions in an era of increasingly capable generative tools. Microscopy already involves substantial processing — colouring, contrast adjustment and reconstruction from grayscale data — so adjudicators must decide where acceptable visualisation ends. Nikon has not yet detailed how it distinguishes permitted processing from prohibited generation, nor whether its procedures will be revised.
For now, the company is standing by its process while promising further word. Dr Xu denies wrongdoing, critics remain unconvinced, and the scientific community awaits a resolution that could shape how imaging contests police authenticity in the future.
(0)