irwan-setiawan
Nikon has launched a review of the winning entry in its prestigious Small World in Motion competition after a wave of criticism from scientists who suspect the video was created or altered using artificial intelligence. The annual contest, now in its fifteenth year, invites researchers and photographers to submit still images and short films captured through powerful microscopes, celebrating the intersection of science and visual artistry. This year's top prize went to Dr Ning Xu, a researcher affiliated with Tsinghua University in China, whose submission purported to show cilia — tiny, hair-like structures — beating inside the airways of a child suffering from a rare disease. The video's striking colours and apparent clarity initially impressed judges, but the celebration quickly gave way to scrutiny.
The backlash began when fellow scientists and former competition judges publicly questioned the authenticity of the footage. Their concerns centred on several anomalies. Edward Phelps, an associate professor at Florida University, noted in comments on Nikon's social media post that some structures in the video appeared to "pop in and out of existence," while the cilia seemed to "appear from nowhere." Ian Donovan, a PhD student at UT Southwestern Medical Centre, pointed out that the source image carried a distinctive watermark typically associated with AI-generated content. Two former judges of the competition also voiced doubts about whether the microscopic recording was genuine. These critiques collectively suggested that the video might not represent a real biological sample observed under a microscope.
Dr Xu has denied breaking any competition rules. While he did not respond to a request for comment from the BBC, he has posted online defending his work. In a reply to Nikon's LinkedIn announcement of his win, he acknowledged using AI but insisted it was not for generating the experimental movie, the cilia, or their motion. He explained that AI was employed "to distinguish and visualize features in the reconstructed grayscale images," a step he considered permissible. Xu also noted that the vivid, artificially bright colours in his video — a common technique in microscopy achieved through stains and dyes applied to samples — were within the contest's guidelines. He maintained that his use of AI complied with the rules, which explicitly forbid generative AI from producing the content but allow image processing tools.
The controversy has thrust Nikon into an uncomfortable position. The company stated that Xu is "respectfully complying" with its investigation and has supplied additional technical documents to support his case. So far, Nikon has publicly stood by its procedures. "At this time, we don't see that any rules were violated," a statement from the company read, though it promised an update soon. The outcome of the review could have significant implications for the competition's credibility, as well as for how scientific imaging contests handle the growing presence of AI in research workflows.
The stakes extend beyond the $3,000 prize. Dr Patrick Hickey, who placed fifth in the 2026 Small World contest, emphasised that the recognition itself carries far more weight than the monetary award. "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. Hickey described the competition as a unique fusion of science and art, where many images possess abstract and artistic qualities. For researchers who labour in relative obscurity, the contest offers a rare platform to share the hidden beauty of the microscopic world. That makes any suggestion of foul play particularly damaging, as it could undermine trust in the winners and discourage honest participation.
Hickey also raised concerns about fairness. He pointed out that the competition has a specific set of rules, one of which clearly prohibits using generative AI to produce entries, while another requires that all footage be taken under a microscope. If the winning video was unfairly altered by AI, he argued, it would be unjust to the other entrants who adhered strictly to those conditions. His comments reflect a broader anxiety within the scientific community: as AI tools become more accessible and sophisticated, distinguishing genuine experimental data from AI-assisted fabrications is increasingly difficult.
The Nikon case is not an isolated incident. It highlights a widening debate about the role of AI in scientific publishing and imaging. AI algorithms are already used to enhance, denoise, and segment microscopic images, which can be legitimate and helpful. However, generative AI models can also fabricate entirely realistic-looking biological structures, raising the risk of fraud. Many journals and competitions have rushed to update their policies, but enforcement remains a challenge. The Small World in Motion contest, with its emphasis on visual appeal, may be especially vulnerable to manipulation because aesthetic quality can sometimes overshadow technical verifiability.
Xu's defence hinges on a distinction between generative AI and AI-assisted visualisation. He claims the underlying data — the grayscale images — were real and that AI simply helped make features more distinguishable. But his critics argue that if AI reconstructed or interpolated missing details, the resulting video might not faithfully represent the original specimen. The watermark detected by Donovan suggests that at least part of the process involved a generative model, which would violate the spirit, if not the letter, of the rules. Nikon's investigation will need to determine exactly how the video was produced and whether the use of AI crossed the line from acceptable processing to fabrication.
The incident also raises questions about transparency in scientific competitions. Should entrants be required to submit raw, unprocessed data alongside their polished submissions? Should they disclose every AI tool used and describe its role in detail? Some scientists are calling for mandatory provenance metadata, similar to the standards emerging in photography and journalism. Without such measures, they warn, contests risk becoming showcases for AI-generated illusions rather than genuine scientific observation.
For now, the microscopy community awaits Nikon's final verdict. The company's decision could set a precedent for how similar disputes are handled in the future. If Nikon clears Xu, it may need to clarify its rules to permit a broader range of AI-assisted techniques, which could alienate purists. If it disqualifies him, it would send a strong message that generative AI has no place in a competition meant to honour authentic microscopic imaging. Either way, the episode has already sparked a necessary conversation about integrity, technology, and the boundaries of scientific visualisation.
Dr Hickey's sentiment likely resonates with many of his peers: the contest is about sharing the wonder of the unseen world, not about winning at any cost. "There's an interesting fusion between science and art," he said, "and a lot of the images that scientists are making have some very abstract and artistic qualities to them." That fusion depends on trust — trust that what viewers see is a real window into biology, not a digital fabrication. As AI continues to blur the lines between reality and simulation, competitions like Small World in Motion will need to adapt to preserve that trust. Nikon's next move will be watched closely by scientists, photographers, and AI ethicists alike.
Comments (0)