Summary
Nikon is reviewing the winning video in its Small World in Motion microscopy contest after scientists alleged that AI post-processing introduced biologically implausible structures. Fredzone reports that creator Ning Xu, an optical engineer at the National University of Singapore, acknowledged using AI on reconstructed grayscale data to distinguish and color similar forms, while maintaining that the underlying lung-cilia movement was real. Critics cited static nucleus-like shapes, impossible mitochondria-like forms and a visible Google SynthID watermark as reasons to question what was recorded and what was generated. Nikon’s rules reportedly prohibit AI image generation, and the company is examining Xu’s technical documentation before deciding whether the entry violated them.
Hana’s perspective
I think this dispute exposes a boundary that creative media also needs to draw clearly. AI-assisted color, cleanup and reconstruction can make images easier to understand, but once a model invents plausible-looking structures, the result is no longer a transparent record. In entertainment that may be an aesthetic choice; in scientific imaging it can alter the evidence itself. Disclosure alone is not always enough if viewers cannot distinguish measurement from synthesis. Competitions, documentary teams and educational producers should require access to raw material, label generated regions and define permitted transformations before judging begins. Trust depends on a visible chain from captured data to finished image.
Source: Fredzone, “AI Distorts Science: The Nikon Small World 2026 Contest Scandal,” https://www.fredzone.org/lia-fausse-la-science-le-scandale-du-concours-nikon-small-world-2026/
Are you a writer with a great story? We'd love to hear from you: APPLY HERE