Nikon Small World in Motion winner faces backlash over claims the video looks AI-made

Started by Drogba, Oct 02, 2026, 08:21 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Topic: Nikon Small World in Motion winner faces backlash over claims the video looks AI-made   Views(Read 116 times)

Drogba

There is a row brewing in the science photography world. The winning entry in Nikon's Small World in Motion competition, which celebrates videos taken through microscopes, has been criticised by scientists who believe it may contain artificial elements. The video, by Dr Ning Xu of Tsinghua University, shows the tiny hair-like structures called cilia moving inside the airways of a child with a rare disease. It is the kind of image that normally wins praise for revealing hidden biology

Several scientists have raised concerns. They say structures appear and disappear in ways that do not match real biology, the cilia look larger than they should, and there are what look like suspicious watermarks that might suggest AI generation. Researcher Edward Phelps said structures seem to pop in and out of existence and cilia appear from nowhere. Dr Patrick Hickey pointed out that the competition rules explicitly ban generative AI and require images to be authentic microscope captures. Some of the criticism has been fairly pointed

Dr Xu has denied that AI was used to generate the video, the cilia or their movement. He did say AI was used for distinguishing and visualising features, which is common in scientific image processing. Nikon said that, at this time, it does not see that any rules were broken, and that Xu is cooperating with its investigation. The company has promised an update soon

This shows how blurry the line has become. Scientists use AI tools all the time to clean up noise, sharpen detail and highlight structures in microscope images. That is different from generating content that was never there, but from the outside the two can look similar. Competitions now have to work out where enhancement ends and fabrication begins. Even experts can struggle to tell the difference on a phone screen. That makes public debates messy

I do not want to prejudge the investigation, and it is entirely possible the video is authentic with heavy processing. But the doubts alone show how much trust has been eroded by AI images. Where should the line be for science imagery? And should competitions require the raw footage alongside entries?


DriftHarbour

The sad part is that the subject matters. Studying cilia in children with rare diseases is important medical research. Whatever happens, the attention has shifted from the science to the scandal

Hopefully the research itself is not tainted by association. Rare disease research struggles for attention as it is. This is not the kind it needed

Neon Harper

Small World is one of the most respected science image competitions in the world. A scandal like this could damage its reputation badly

They need to be thorough and transparent with the outcome. Publishing the evidence would help

Past winners must be watching this closely. Their own work could be questioned next

Cass87

Requiring raw footage alongside entries seems like the obvious fix. Judges could compare the original capture with the processed version

That would settle most disputes quickly. It adds a bit of paperwork, but credibility is worth it

Sequence87

This is why science image integrity has become such a big topic in journals. Fake western blots and manipulated microscope images have caused real retractions

AI makes manipulation easier and harder to detect. Competitions are just the most visible part of a bigger problem

William56

Structures popping in and out of existence is a classic sign of generative video. Real cells do not do that

If experts in airway biology are raising this, it deserves a proper look. Nikon cannot just wave it away
Currently losing at something

WCWFinley65

Nikon's statement that no rules appear to have been broken feels premature. Investigate first, then comment

Defending the winner before checking makes them look partial. First impressions count in a row like this. Nikon has handled the start badly

Paul73

AI processing for denoising is completely standard in microscopy now. Almost every modern image goes through some kind of algorithm. The question is whether it invents detail or reveals detail

That distinction matters enormously. Most viewers would never know the difference

ScarletWrench

Fairness to Dr Xu matters too. Public accusations can ruin a career even if they turn out to be wrong. Wait for the investigation before deciding

Plenty of real images look strange to people outside a specialism. Due process matters even online

Amber79

Watermarks are a strange detail. Why would a real microscope video have watermarks that look like an AI generator's? Hopefully the investigation explains that. That detail alone needs a clear answer. Simple questions deserve simple answers