Small World in Motion: After AI Disqualification, Experts Discuss How Striking Microscopy Images Can Stay True to Reality

EDINBURGH, UK – 16 October 2026: Nikon’s decision to disqualify its 2026 Small World in Motion winner over generative AI has sparked a global debate about AI in scientific imaging. On 14 October, Bitesize Bio brought together four experts, including two of the scientists behind the open letter that challenged the video, to ask what counts as acceptable AI use in microscopy and how to make striking images that stay true to the raw data. The recording is now free to watch on demand[1].

The Disqualified Movie that Sparked the Discussion

The winning entry to Nikon’s 2026 Small World in Motion competition was a video captioned “abnormal beating of airway cilia in a child with primary ciliary dyskinesia (PCD)”[2,3]. But shortly after the winner was announced, cilia biologists, microscopists, and PCD clinicians questioned its origins[4].

On 9 October, Nikon ruled that the video “did not comply with the competition rules regarding generative AI” and revised the results, adding that the decision was not a judgment of the entrant’s reputation or intent[5,6].

Five days later, Bitesize Bio hosted a panel discussion to ask a wider question sparked by the debate: with AI tools increasingly used in imaging, what does that mean for scientific data integrity? Within two days of the panel being announced[7], 495 people registered interest, with 243 attending the live webinar and posting 339 comments and questions during the hour.

Highlights from the Panel Discussion

Professor Peter O’Toole (University of York; Vice President, Royal Microscopical Society) chaired the panel. He was joined by PCD researcher Professor Pleasantine Mill (MRC Human Genetics Unit, University of Edinburgh), who led an open letter to Nikon[8] signed by 469 people from 36 countries and territories, from microscopists and clinicians to patients and families[9]; live microscopy expert Dr Simon Cleary (King’s College London), one of the letter’s lead signatories[8]; and Dr Florian Jug (Human Technopole, Milan), creator of the widely used CARE and Noise2Void denoising tools[10].

The panel drew a clear line between enhancing and generating light microscopy images. Cleary said, “The line is crossed when you make images that don’t represent the positions of real photons in the world. That’s generation, not bringing out real signal.” He suggested that scientific images must be based on reality and that researchers must declare anything they change.

Mill argued that the incident exposed a training gap, not ill intent: “It’s about culture. A generation is coming through for whom everything is filtered by default: the photos on our phones, the posts we share. So it’s not surprising that trainees don’t always see why this matters for what we acquire on a microscope, why raw data matters. More scripted, hardwired guidelines will only put people off. What we need is upskilling, so that everyone understands why authenticity in their data is essential.”

Jug pointed out that the problem itself is not new: “Faking results to claim a finding you didn’t make is not a new problem. The scientific record has its share of people who had wonderful careers, until they didn’t, built on fabricated data. What’s new is that it’s now possible with microscopy images, and it’s cheap and accessible. It no longer takes a team led by someone with bad intentions. Anyone under enough pressure could do it, at any career stage. What worries me more is when a PI unknowingly believes something has been found when it hasn’t. So we need to be more on our toes.”

On striking a balance between beauty and fact, Jug said, “I’d love to keep the attraction of a beautiful image, with a small link beside it where you can download the raw TIFF: noisy, disgusting and real. That would be super cool.”

Next Steps for Science and Competitions

The panel noted that the imaging community needs to keep this conversation going. They agreed on the following suggestions for scientists:

• Publish raw images and video files alongside processed versions
• Declare clearly how images have been processed
• Train researchers how to correctly handle image data

And on approaches to safeguard imaging competitions:

• Have domain experts review shortlisted entries for authenticity
• Share shortlisted images before choosing winners to allow for public scrutiny

The panel also discussed watermarking AI-generated images, but did not reach agreement.

Dr Nick Oswald, Founder and Director of Bitesize Bio, said: “Microscopy images are data, and researchers need to be able to trust them. As AI tools become more common in imaging, researchers need to think carefully about when they’re the right tool and be open about how they’ve used them. This panel was a starting point, and we’ll keep making space for the community to work out what good practice looks like.”

View the full on-demand recording here.

Sources

1. Bitesize Bio, Nikon Small World and AI: Where Should Scientific Imaging Draw the Line? (on-demand recording), 14 October 2026. All panel quotes are drawn from this recording and approved by the speakers.
2. Nikon Small World, 16th Annual Nikon Small World in Motion Competition winner announcement, PR Newswire, 16 September 2026.
3. Nikon Small World, Abnormal beating of airway cilia from a child with Primary Ciliary Dyskinesia (PCD), competition gallery entry.
4. CNN via KEYT, Winner of Nikon microscopy photo contest faces criticism over AI usage, 1 October 2026.
5. The Scientist, Nikon disqualifies winning video over AI use following weeks of controversy, October 2026.
6. BBC News, report on Nikon’s ruling, 9 October 2026.
7. Bitesize Bio, panel announcement, 13 October 2026.
8. Pleasantine Mill, Please sign and share: open letter to Nikon on the 2026 Small World in Motion winner, FocalPlane (The Company of Biologists), 8 October 2026.
9. FocalPlane, Update on the open letter to Nikon Small World in Motion, 14 October 2026.
10. Human Technopole, Florian Jug profile.

Notes to editors

About Bitesize Bio (Science Squared Ltd): Bitesize Bio publishes practical articles, webinars, and podcasts for life science researchers, including The Microscopists podcast hosted by Peter O’Toole. Visit bitesizebio.com

About the panelists:

• Professor Peter O’Toole is Director of the Bioscience Technology Facility and Head of Imaging and Cytometry at the University of York, and Vice President of the Royal Microscopical Society.
• Professor Pleasantine Mill is a Professor of Cilia Biology and MRC Investigator at the University of Edinburgh’s MRC Human Genetics Unit, and Co-Director of the LifeArc Centre for Rare Respiratory Diseases. She coordinated the open letter to Nikon.
• Dr Simon Cleary is a Lecturer at King’s College London’s Institute of Pharmaceutical Science and Centre for Lung Health, with experience in video microscopy.
• Dr Florian Jug is Head of Research, Multimodal AI Across Scales, at Human Technopole in Milan, Italy.

Press kit available here: https://drive.google.com/drive/folders/1WPo1yOwL7hBAVoUnBc5xOf4IZxb7lIg6?usp=drive_link

Speakers are available for interviews through Bitesize Bio’s media contact: Zara Puckrin, Marketing Manager, zara@bitesizebio.com, +44 (0)7593 699531.