Full Breakdown
Nikon Strips First-Place Video After AI-Rule Violation Review
By Drooid · · How we work
Core Event: Disqualification of the 2026 Small World in Motion Winner
On October 9, 2026, Nikon announced that the video submitted by Dr Ning Xu was removed from first-place standing in its Small World in Motion microscopy competition. The decision altered the rankings, promoting Nguyen Nam Nhat of Vietnam to first place.
Background & Context
The Small World program began in 1975 as a still-photo contest and added a moving-image category in 2011. The 2026 edition attracted 347 entries from 40 countries, reflecting the growing popularity of high-resolution microscopy. In recent years, generative-AI tools have become common for post-processing scientific images, prompting Nikon to prohibit fully AI-generated videos while allowing conventional image enhancements.
Data & Statistics
- Entries: 347 submissions (2026) – 40 nations (contest organizers).
- Prize: First-place award carries a US $3,000 prize (DPReview).
- Original winner: Dr Ning Xu, listed as affiliated with Tsinghua University in China by the BBC and The Scientist; other outlets identified him with the National University of Singapore.
- New winner: Nguyen Nam Nhat, Vietnam, video of a roundworm and the single-celled organism *Dileptus*.
Official Statements & Responses
- Dr Ning Xu: In a LinkedIn comment, Xu described using an “unsupervised neural-network method” for “AI-assisted post-processing” to colour features in reconstructed grayscale images. He asserted that AI was not used to generate the experimental footage, the cilia, or their motion, and that he believed his work complied with the rules.
- Competition spokesperson (Remi Yuter, quoted by Gizmodo): The disqualification rests on rule eligibility, not on any assessment of Xu’s reputation or scientific intent.
Criticism & Opposition
- Edward Phelps, University of Florida: Noted that the video displayed structures that appeared and vanished without continuity, a pattern he said is characteristic of AI-generated content.
- Melanie White, University of Queensland: Warned that scientific images must be grounded in actual measurements, emphasizing the need for trust in microscopy data.
On-the-Ground Reports
Scientists and former judges posted concerns on Nikon’s LinkedIn page shortly after the September winner announcement. A doctoral student, Ian Donovan of UT Southwestern Medical Center, identified a SynthID watermark—a marker Google DeepMind embeds in AI-generated media—in the video file. While the watermark was reported by several outlets, Nikon’s statement did not cite it as a basis for the disqualification.
Conflicting Reports & Gaps
- Affiliation of Dr Ning Xu: The BBC and The Scientist list him as a researcher at Tsinghua University, China; other sources (PetaPixel, Nature) cite the National University of Singapore.
- Extent of AI use: Xu claims AI was limited to colourisation after reconstruction, whereas critics argue that generative networks may have altered pixel data beyond simple colour mapping. No source provides the specific neural-network architecture, parameters, or raw data to verify the boundary.
- SynthID watermark: Some outlets report its presence; Nikon’s official note makes no reference to a watermark, leaving its role unverified.
Verbatim Quotes
- “I see myself as a ranger in the African bush,” — Benedikt Pleyer, third-place winner.
What’s Next
Nikon has pledged to revisit its competition rules and evaluation procedures, acknowledging the difficulty of distinguishing AI-assisted processing from fully AI-generated content in scientific microscopy. The organization has not announced a timeline for revised guidelines or any additional verification requirements for future entrants.