In brief
Nikon’s annual microscopic photography contest has returned to the spotlight when the 2026 winner, a vivid video of human cilia, was accused by peers of using AI to fabricate details. Dr Ning Xu, the Chinese researcher who won, denies the claim, saying he only used AI for image enhancement and stayed within contest rules. Critics point to unnatural shapes and movements, while Nikon maintains that no rule was breached, yet an investigation remains in progress. Former judges, a PhD student, and other experts weigh in, highlighting the growing tension between cutting‑edge imaging techniques and the ethical use of AI in scientific research. The outcome will influence future contest guidelines and scientist‑public trust.
The winner of Nikon’s 15‑year‑old Small World In Motion contest is being re‑examined after a group of scientists raised doubts that the winning video was artificially altered with AI.
The contest invites amateurs and professionals to submit videos and photographs taken with advanced microscopes, rewarding the best with a small cash prize and recognition within the scientific community.
The winning entry, submitted by Dr Ning Xu of Tsinghua University in China, shows a close‑up of hair‑like cilia moving inside the airway tissues of a child suffering from a rare respiratory condition. The footage is brightly coloured – a standard practice in microscopy achieved through staining – and contains seconds of cilia beating in a pattern that appears almost cinematic.
When queried by the BBC, Xu denied that the image was fabricated, saying: “AI was not used to generate the experimental movie, the cilia, or their motion. I only employed AI to highlight and visualise features in the grey‑scale reconstructions, which is allowed under the contest rules.” He added that any editing remained within the established guidelines.
However, a number of fellow scientists criticised the video. Educationist Edward Phelps of Florida University said certain structures seemed to “pop in and out of existence,” while PhD student Ian Donovan of UT Southwestern Medical Centre pointed out a watermark that suggests deep‑learning synthesis. Two former judges also expressed doubts regarding the image’s authenticity.
Nikon has responded that Xu is “respectfully complying” with its inquiry and has supplied additional technical documentation. In a statement, the company emphasised that “none of the contest rules appear to have been violated” at present and that an update will follow after the investigation concludes.
Dr Patrick Hickey, who placed fifth in the 2026 contest, commented on the implications: “The award carries more than a small monetary prize – it recognises the intersection of scientific rigor and artistic expression. If it turns out that the winning entry was manipulated unfairly, it would unfairly disadvantage deserving entrants.” He reiterated that the contest rules clearly disallow generative AI to create content.
The debate underlines the tension growing within scientific communities as AI tools become more pervasive. The outcome of Nikon’s review will likely inform the adjudication standards for future contests and shape expectations about the ethical use of AI in research imagery.
