The Daily Pulse

Science & Environment

Nikon investigates top microscopy prize winner over AI concerns

Reported by Rituparna Sen (Section Editor) · BBC News - Science & Environment ·

Verified — Confirmed by a second source: Google News - UK Science

Confirmed by a second source: Google News - UK Science

Important
Nikon investigates top microscopy prize winner over AI concernsRepresentative image · Pexels

Nikon is checking the winning entry of its yearly microscope video contest after other scientists claimed the footage used artificial intelligence. The top spot went to a short clip sent in by Tsinghua University researcher Dr Ning Xu, which purported to show tiny hair-like cilia moving inside a child's sick airway.

The Dispute

Xu has not replied to direct interview requests but stated online that he did not break any contest guidelines. While he admitted to using artificial intelligence tools to help show features in grayscale frames, he maintained that the actual movie and the moving cilia were not made by computer software. Other experts pointed out strange shapes and movements that looked unnatural, with Florida University associate professor Edward Phelps noting that certain structures appeared and vanished unexpectedly. UT Southwestern Medical Centre PhD student Ian Donovan also spotted what looked like an artificial generation watermark, while two former contest judges shared similar doubts.

Rules and Reaction

Nikon stated that Xu is cooperating with the review and has shared extra technical papers, though the company currently finds no rule breaches and will share an update soon. Contest fifth-place finisher Dr Patrick Hickey noted that the prestige of the award matters more than the three thousand dollar cash prize, adding that unfair artificial alterations would cheat other participants.

Context

Scientific photography contests often struggle to set clear boundaries as image-enhancing software becomes more advanced and widely available. These competitions typically allow basic color adjustments and contrast work to make hard-to-see samples visible to the human eye, but draw a strict line against using computer software to invent content that was not captured through the lens.

Read more at BBC News - Science & Environment →