In the quiet discipline of microscopy, where scientists spend long hours in darkened rooms coaxing meaning from the invisible, a question has surfaced that reaches far beyond one competition: when does a tool for seeing become a tool for inventing? Nikon's prestigious Small World In Motion contest is now investigating its winning entry — a video of cilia moving through a child's airways, submitted by Dr Ning Xu of Tsinghua University — after fellow scientists raised doubts about whether artificial intelligence did more than assist in the work, potentially crossing the boundary from enhancement
Nikon photo contest winner faces AI scrutiny as scientists question microscopy video
AI was not used to generate the cilia, or their motion
Mark
So the question is whether Xu used AI to enhance real microscope footage, or whether he used it to generate the footage itself?
Mimi
Exactly. He admits to the first—using AI to process and highlight features in grayscale images. But other scientists are saying the structures move in ways that don't match biology, and that some appear and disappear, which suggests generation rather than enhancement.
Luke
But we should be careful here. Phelps and Donovan are raising concerns, not proving fraud. A watermark *suggests* AI generation, but watermarks can appear in legitimate images too. We don't have Xu's raw data or Nikon's full analysis yet.
Mark
What's the actual rule in the contest?
Mimi
No generative AI to produce things. The footage has to come from a microscope. Xu's claim is that the cilia and their motion came from the microscope—he just used AI to make them visible.
Luke
Which is a meaningful distinction if true. But it's also exactly the kind of distinction that's hard to verify without seeing the original files and the processing steps.
Mark
Why does this matter beyond the competition itself?
Mimi
Because it's about trust in scientific imagery. These images are used in research, in publications, in understanding disease. If people can't tell what's real and what's been artificially enhanced or generated, the whole system breaks down.
Luke
And because it raises a real question about where the boundary should be. Is AI-assisted visualization of real structures acceptable? Most scientists would say yes. Is AI-generated structure presented as real unacceptable? Yes. But the space between those two is getting crowded.
Mark
What happens now?
Mimi
Nikon is investigating. They've said no rules were violated yet, but they're promising an update. The scientific community is watching closely.
Luke
And Xu hasn't directly engaged with the specific technical criticisms. He's said he followed the rules, but he hasn't addressed whether the cilia really move that way, or why structures appear and disappear.
The Pulse
- Scientists watching the winning video noticed cilia appearing and vanishing in ways that defy known biological behavior, raising alarms that something other than a microscope captured the footage.
- A PhD student identified what appears to be a generative AI watermark embedded in the source image, adding a technical dimension to what began as an intuitive unease among competitors.
- Dr Xu has defended himself online, acknowledging AI was used to highlight features in grayscale images but insisting it did not create the cilia, their motion, or the underlying footage.
- Two former competition judges have joined active researchers in voicing concern, amplifying pressure on Nikon to move beyond its initial statement that no rules appear to have been broken.
- Nikon has opened a formal investigation and is collecting technical documentation from Xu, but has yet to reach a conclusion — leaving the scientific imaging community in an uncomfortable suspension.
In the quiet discipline of microscopy, where scientists spend long hours in darkened rooms coaxing meaning from the invisible, a question has surfaced that reaches far beyond one competition: when does a tool for seeing become a tool for inventing? Nikon's prestigious Small World In Motion contest is now investigating its winning entry — a video of cilia moving through a child's airways, submitted by Dr Ning Xu of Tsinghua University — after fellow scientists raised doubts about whether artificial intelligence did more than assist in the work, potentially crossing the boundary from enhancement into fabrication. The dispute arrives at a moment when the scientific community is still learning to draw lines that technology keeps quietly redrawing.
Dr Ning Xu's winning entry in Nikon's Small World In Motion competition — a video of hair-like cellular structures called cilia moving through a child's airways — has drawn sharp scrutiny from scientists who believe artificial intelligence played a role in its creation beyond what the contest permits.
The Small World In Motion competition, which Nikon has run for fifteen years, is among the most prestigious awards in scientific imaging. Its $3,000 prize matters less than the recognition it offers researchers who spend their working lives peering into microscopic worlds. Xu, based at Tsinghua University in China, did not respond directly to BBC inquiries but posted a public defense online, denying any violation of the rules. He acknowledged using AI "to distinguish and visualize features in the reconstructed grayscale images," while insisting that "AI was not used to generate the experimental movie, the cilia, or their motion."
Other scientists have challenged that distinction with specific observations. Edward Phelps, an associate professor at Florida University, noted that structures in the video "pop in and out of existence" in ways inconsistent with real cellular behavior. Ian Donovan, a PhD student at UT Southwestern Medical Centre, identified what he described as a watermark typically associated with AI-generated imagery. Two former competition judges have also expressed doubts about the video's authenticity, pointing to cilia that appear disproportionately large and movements that do not align with known biology.
For those who placed below Xu, the stakes are personal as well as principled. Dr Patrick Hickey, who finished fifth, described the competition as a rare chance for scientists to share their work with the public — and argued that if the winning entry was produced with prohibited tools, it would be a straightforward injustice to those who followed the rules.
Nikon has launched an investigation and says Xu is cooperating by providing technical documentation. The company has not yet concluded that any rules were broken, but has promised an update. The case has laid bare a tension that will only deepen: as AI becomes more woven into scientific workflows, the difference between seeing something and constructing it grows harder to define — and harder to police.
Dr Ning Xu's winning entry in Nikon's Small World In Motion competition—a short video of hair-like cellular structures called cilia moving through the airways of a child with a rare disease—has become the subject of intense scrutiny from other scientists who believe artificial intelligence played a role in its creation beyond what the contest permits.
The Small World In Motion competition, which Nikon has run annually for fifteen years, invites researchers to submit photographs and videos captured through powerful microscopes. It is among the most prestigious awards in scientific imaging, valued not primarily for its $3,000 prize but for the recognition it confers on those who spend their working hours peering into microscopic worlds. Xu, a researcher at Tsinghua University in China, submitted the winning video this year. When asked by the BBC to comment, he did not respond directly, but he did post a response online to criticism raised on Nikon's LinkedIn page, denying that he had violated the competition's rules.
The core dispute centers on what constitutes permissible image processing versus prohibited artificial generation. Xu's video displays the vivid, artificial coloring common in microscopy work—typically achieved by applying stains and dyes to samples, though it can also be done computationally. This kind of enhancement is allowed under contest rules. Xu acknowledged using AI "to distinguish and visualize features in the reconstructed grayscale images," but he insisted that "AI was not used to generate the experimental movie, the cilia, or their motion." The distinction he draws is between using AI as a tool to process and highlight structures that were actually present in the original microscope footage, versus using AI to create or substantially fabricate the biological material itself.
Other scientists have raised specific objections that suggest the video may cross that line. Edward Phelps, an associate professor at Florida University, noted that some structures in the video "pop in and out of existence" and that cilia "appear from nowhere"—behaviors inconsistent with how actual cellular structures move. Ian Donovan, a PhD student at UT Southwestern Medical Centre, identified what he described as a distinctive watermark in the source image that typically indicates AI generation. Two former judges of the competition have also expressed concerns about the video's authenticity. These challenges point to movements and shapes that do not align with known biological behavior, and to the cilia appearing disproportionately large relative to normal cell structures.
Dr Patrick Hickey, who placed fifth in this year's competition, articulated what is at stake. The award matters to scientists not because of the money but because it offers a rare public platform for work that is ordinarily conducted in isolation. "If you spend a lot of your time in a dark room looking down a microscope, it's really nice to be able to show people and to show the public," he said. He emphasized that the competition has explicit rules against using generative AI to produce imagery, and that footage must originate from actual microscope observation. If Xu's video violated these standards, Hickey argued, it would be unfair to the other entrants who adhered to them.
Nikon has launched an investigation and stated that Xu is "respectfully complying" by providing additional technical documentation. The company has not yet concluded that rules were broken, issuing a statement that "at this time, we don't see that any rules were violated." But it has promised an update soon. The case has exposed a growing tension in scientific work: as AI tools become more sophisticated and more integrated into research workflows, the line between legitimate computational enhancement and impermissible artificial generation grows harder to police. What counts as seeing, and what counts as creating, is no longer obvious.
Notable Quotes
If you spend a lot of your time in a dark room looking down a microscope, it's really nice to be able to show people and to show the public.— Dr Patrick Hickey, fifth-place finisher in the 2026 Small World contest
There's a specific set of rules in the contest, and one of them was quite clearly that you couldn't use generative AI to produce things, and the other was obviously that it has to be taken under a microscope.— Dr Patrick Hickey
Want the full story?
Read the original at BBC News ↗