Nikon Disqualifies AI-Assisted Microscopy Video Winner

Why was the Nikon video winner disqualified?

Nikon said Ning Xu’s entry did not comply with the Small World in Motion competition rules regarding generative AI. The company had reviewed the video after questions were raised about whether it was authentic. ¹²

The video was presented as showing cilia, tiny hair-like structures, moving in the airway of a child with a respiratory condition. Nikon did not say in the available reports exactly which part of its rules the entry breached or what specific processing crossed the line. ¹

What did the entrant say AI was used for?

Xu said an unsupervised neural-network method was used for AI-assisted post-processing. In plain terms, this means a computer method helped pick out and display features in reconstructed grayscale images. ¹

Xu denied that AI generated the experimental movie, the cilia, or their movement, according to Ars Technica. That distinction matters: using AI to process or highlight recorded data is not necessarily the same as using it to invent visual content. The reports do not provide enough detail to independently assess the images or the method. ²

How can a contest identify AI-generated work?

This case points to several possible checks, but the reports do not give a complete account of Nikon’s investigation. Online observers questioned the video’s biological appearance, including apparent structural abnormalities and features that seemed to appear and disappear. Some also reported seeing what looked like AI watermarks in source files. These were concerns raised by observers, not findings that Nikon is reported to have confirmed individually. ²

A contest can compare an entry with its rules and ask for underlying material or an explanation of how it was made. The available sources do not say what records Nikon requested, which technical tests it used, or whether any single check determined the decision. A visual oddity or a detector result alone should not be treated as proof of AI use; the reports do not establish a reliable, universal test for identifying AI-generated work.

What does this case show about disclosure rules?

A rule about generative AI may leave important questions for a contest to define: does it cover generating an image, processing recorded data, or both? What must entrants disclose, and how should judges distinguish enhancement from invention? The sources say Nikon is revisiting its competition rules, but do not report the revised wording or when it will take effect. ¹

For entrants, the practical point is to read the rules closely and explain material AI use when the rules require it or when the process could affect how the work is understood. This is a general takeaway, not a statement of Nikon’s current requirements beyond what the reports describe.

Who is the new winner, and what remains unclear?

After Xu’s entry was disqualified, Nikon named Nguyen Nam Nhat of Vietnam the new winner. His video shows a tiny roundworm and a single-celled organism called Dileptus. ²

The reports do not explain the full evidence behind Nikon’s ruling, precisely how the rules applied to Xu’s post-processing, or what changes Nikon plans to make. Those details will matter to future entrants trying to understand where scientific image processing ends and prohibited AI use begins.