Nikon disqualified the winner of its Small World in Motion microscopy video contest on Friday. The company said the video broke the contest's rules on generative AI. Optical engineer Ning Xu won in September with a video of tiny hair-like structures called cilia beating in the airway of a child with a rare genetic disorder. Scientists soon questioned it online, and Nikon started re-reviewing it in late September. Xu has said he used AI only after filming, to separate and color structures. The BBC said it had approached him for comment on the ruling. Nikon said the ruling isn't a judgment of Xu's reputation or intent. Nguyen Nam Nhat of Vietnam now holds first place with a video of a tiny roundworm and a single-celled organism.

The Video Shows Things That Aren't There

Scientists say the video shows things that don't happen in a real airway.

To a PCD specialist, this doesn't look like the disease. Robert Hirst was among the first to object. He's the lead scientist for the NHS's diagnostic centre for the disorder at the University of Leicester. He told the BBC he "knew immediately that the video was fabricated" and wasn't surprised by the ruling. He's diagnosed the disease for 20 years, and he says the cells and cilia look nothing like a patient's.

Critics say parts of the video can't happen in biology. Edward Phelps is an associate professor at the University of Florida. He said purple structures beneath the cilia look like mitochondria the size of cell nuclei, which doesn't occur in biology. He said some shapes pop in and out of existence. A UT Southwestern PhD student, Ian Donovan, said the source image appeared to carry an AI watermark.

A science image has to show real data. Melanie White, a developmental biologist at the University of Queensland, told Nature that viewers need to trust an image comes from the actual measurement. Berlin researcher Valentin Dunsing-Eichenauer put it more bluntly on LinkedIn: what Xu called "visual presentation" is "generating fake."

A rival entrant says the AI rule was clear. Patrick Hickey, who now sits in fourth place, told the BBC that one rule was "quite clearly that you couldn't use generative AI to produce things." He said that if AI had unfairly altered the winner, that would be unfair to everyone else.

But Coloring Data for a Beauty Contest Isn't Faking Science

Xu and some imaging researchers say AI visualization is fine when it doesn't make a scientific claim.

The cilia and their motion are real, Xu says. He wrote on LinkedIn that AI didn't generate the movie, the cilia or their motion. He said AI came in afterward to separate and color structures in grayscale data.

He says the colored parts were never meant as anatomy. Xu wrote that the contest celebrates the beauty of microscopy, so his team made the area under the cilia more visually engaging. He said they made no claims about what those rendered features are. In his view, a scientific claim and a beautiful visualization can both have a place, as long as people know which is which.

One expert says AI visualization isn't automatically wrong. Markus Sauer, a super-resolution microscopy researcher at the University of Würzburg, told Nature that AI imagery becomes a problem when the images misrepresent, exaggerate or alter the findings. Coloring is common in microscopy, and the contest allows it.

Either Way, Nikon's Contest Missed It

Critics say the contest should have caught this before it crowned a winner.

Former judges and other outsiders caught it before the contest did. Scientists raised the alarm in comments on Nikon's own LinkedIn posts. Even after the backlash began, Nikon said it saw no broken rules. On Friday it said the episode showed it needs to revisit its rules and how it evaluates entries.

AI scandals have hit photo contests before. Gavin Stoker of Amateur Photographer pointed to a 2023 Sony World Photography Awards winner that its creator said was made with AI. He also cited a 2026 Hasselblad Masters finalist that online commenters exposed after officials missed it. He wrote that after those cases, you'd assume judges would put shortlisted entries "under the microscope."

One contest director also blames competition culture. Shawn Reeves, who directs a high school physics photo contest, said on LinkedIn that Xu should have withdrawn. He also said "part of the blame lies in the structures and superstructures of competition."

Where This Lands

Hirst, Phelps and other scientists say the video shows structures that don't occur in real biology, and Hirst wasn't surprised Nikon pulled it. Xu says AI only separated and colored structures in real footage. He says his team rendered the area under the cilia for looks, because the contest is about beauty. Sauer says using AI to visualize data isn't automatically a problem unless it distorts the findings. Stoker wrote that after earlier AI scandals, you'd expect judges to check shortlisted entries closely. Reeves blames both Xu and the culture of competition. Nikon now says it will revisit its rules and judging for future entries.

Sources