At the Royal Botanic Gardens, Kew, 180 year old specimens are yielding genetic data, and new tools are giving researchers across Africa, Asia, and Latin America access to archives once locked away.
Inside the Royal Botanic Gardens, Kew, a fungus specimen collected in Madagascar nearly two centuries ago is still giving up its secrets. Recent genetic work on the 180-year-old Pisolithus madagascariensis, a gasteroid fungus endemic to the island, has produced recoverable DNA that Kew scientists describe, in coverage of a new Kew report by The Guardian, as a potential "genomic goldmine" for cataloguing life that has never been formally named.
The specimen is one of millions stored in Kew's herbarium, the global archive of pressed and preserved plants and fungi that underpins most of what scientists know about plant diversity. Kew's State of the World's Plants and Fungi report argues that AI and digitization could turn this archive into something more useful than it has ever been. "We are in a race against extinction," Alexandre Antonelli, Kew's director of science, told The Guardian. The technology, he said, is finally catching up to the crisis.
The numbers in the report are stark. About 40% of the world's roughly 70,000 assessed plant species are at risk of extinction. Somewhere around 90% of fungal species remain unknown to science. Kew estimates that researchers name only about 2,000 new plant species a year, a pace Antonelli calls a scratch on the surface of the work remaining.
New tools are changing what is possible. High-resolution imaging and machine learning can identify specimens far faster than specialists working by eye, according to the Kew findings reported by The Guardian. Kew researchers say the systems can match or beat trained taxonomists on difficult groups such as sedges and peat mosses, plants that have long resisted identification because their distinguishing features are microscopic. Online access to digitized collections, meanwhile, has opened Kew's archive to botanists in Africa, Asia, and Latin America who previously had to travel to London to consult it.
The push matters because plants and fungi quietly underpin much of what humans depend on. Food crops, medicines, the carbon storage of forests, and the cycling of nutrients through soil all rest on species that have, in many cases, never been named. Losing a species before it is described means losing the chance to learn what it did.
The Kew report frames AI as a turning point rather than a solution. Antonelli's own caveat, that the 2,000-species-a-year naming rate would have to multiply many times over to close the gap, runs through the coverage. The risk is that enthusiasm for AI tools softens the loss rather than slowing it. The constructive case is narrower than the headlines suggest: digitization is helping scientists see the crisis more clearly, and it is letting researchers outside the traditional centers of botanical power join the work. Whether that is fast enough remains the open question.