Nikon Strips Small World in Motion Winner Over AI Use

Nikon has disqualified the winning entry of its 2026 Small World in Motion video competition after an investigation found the video relied on AI in ways that broke contest rules. The decision, announced following an October update, strips first place from Ning Xu of Tsinghua University in China and hands the title to Nguyen Nam Nhat of Vietnam, whose video shows a tiny roundworm interacting with a single-celled organism.

Xu's original entry had drawn attention last month for its subject: tiny cilia beating in the airways of a child with a rare respiratory disorder. The footage appeared to offer a rare microscopic view of a medically significant phenomenon. Then researchers began raising questions.

According to a BBC report, some researchers noted structural abnormalities in the video and observed that cell features appeared and disappeared in ways that do not occur in actual cellular biology. Others found what looked like AI watermarks embedded within the source files. Nikon launched an investigation and concluded that Xu's entry did not comply with competition rules.

What did the disqualified researcher say?

Xu acknowledged using AI tools but framed their role narrowly. He said he used AI "to distinguish and visualize features in the reconstructed grayscale images." He denied that AI was used "to generate the experimental movie, the cilia, or their motion."

That distinction did not satisfy the competition's judges. Nikon's ruling effectively defines the line the contest draws: post-processing assistance may be acceptable, but AI involvement that shapes what viewers see as biological reality crosses the rules.

The case matters beyond one contest. Scientific imaging competitions, journals, and publishers are all grappling with where AI assistance ends and fabrication begins. Microscopy has long depended on image enhancement—contrast adjustment, deconvolution, reconstruction—but generative tools blur the boundary between revealing what a microscope captured and inventing what it did not. A disqualification at a competition of this stature signals that organizers now treat AI provenance as a checkable rule, not an honor-system footnote.

What is the Small World in Motion competition?

Nikon founded the Small World in Motion competition in 2011 as a spinoff of its Small World Photomicrography Competition, which dates to 1974. Both contests aim, in Nikon's words, "to showcase the beauty and complexity of things seen through the light microscope"—one through still photographs, the other through video and time-lapse imaging made possible by decades of technological advance in microscope-mounted cameras.

The competition has become one of the most visible showcases for photovideography in science. Winning entries circulate widely in press coverage and classrooms. That visibility raises the stakes for authenticity: a fabricated or AI-altered winner would mislead not just judges but the public's understanding of what cells and organisms actually look like.

Who are the new winners?

With Xu's disqualification, every remaining winner moved up one place:

• First place: Nguyen Nam Nhat of Vietnam, for a video of a tiny roundworm and a single-celled organism.
• Second place: Benedikt Pleyer of Bavaria, Germany, for a video of jellyfish larvae suspended in water droplets.
• Third place: Andrew Moore of Howard Hughes Medical Institute in the US, for a video of synchronized cell division in neighboring cells.
• Fourth place: Patrick Hickey of Hypha Research Ltd. in Edinburgh, Scotland, for a video of the dynamics of mitochondria and chloroplasts in Turtle Vine leaf cells.
• Fifth place: Igor Adameyko, for a video of a tiny, colorful sea slug (Cyerce kikutarobabai) on the move.

The reshuffled results still span the range the competition exists to celebrate: developmental biology, cell division, organelle dynamics, and marine microfauna, all captured through the light microscope without generative assistance.

What does this mean for AI in scientific imaging?

The disqualification lands amid a broader reckoning over AI in research imagery. Detection in this case came from the community, not from contest screening: researchers flagged biological implausibilities, and watermark artifacts in the source files supplied concrete evidence. That pattern suggests competitions and journals may need formal provenance checks—raw file submission, processing logs, or AI-use disclosure—as AI tools become standard parts of the imaging workflow.

Xu's defense also previews the hard cases ahead. If AI is used "to distinguish and visualize features in the reconstructed grayscale images," as he put it, the output may be genuinely helpful for analysis while still failing a competition built on authentic photovideography. Nikon's ruling indicates that for this contest, at least, authenticity of the visible image is the standard—and future entrants now know that AI watermarks in source files can cost them the title.