The writer is a science commentator
Confirming whether a plant or fungus is extinct is not a cut-and-dried issue. If a fern or orchid known to grow in one place has not been observed on repeat visits, has it really disappeared forever? Might it be flowering, unobserved, at a higher altitude because of climate change? Fungal networks can thrive, unseen, under the soil.
The dilemma over when to declare a plant extinct is highlighted in a new report, published this week by the Royal Botanic Gardens, Kew, showing how digitisation and AI are changing conservation. While the report reveals that the real scale of species loss is underestimated, it is also true that plants once thought extinct, such as the Chilean blue crocus, have cropped up again after concerted efforts to find them.
“It’s very difficult to prove that something is no longer there,” Alexandre Antonelli, executive director of science at Kew, told me on Monday at the launch of the 2026 State of the World’s Plants and Fungi report. That is why scientists are moving away from yes-no verdicts and turning to mathematical models that calculate the probability that a plant still exists. The digitisation of Kew’s 7.4mn plant and fungi collection, plus the use of AI, is helping with the shift, by linking specimens to searchable information that has a bearing on survival.
Finding smarter ways to protect biodiversity matters: the rich web of species is both a support system for all life on Earth and a backbone of the economy. It helps to maintain, among other things, global water supply, food security and the flow of natural commodities. Leaving aside the moral case for saving species, rescuing plants and fungi on the brink, rather than hunting down those already lost, could also ultimately lead to valuable new compounds, such as medicines or plastic-digesting substances.
The International Union for Conservation of Nature defines a species as “presumed extinct in the wild when exhaustive surveys in known and/or expected habitat, at appropriate times . . . throughout its historic range, have failed to record an individual”. Exhaustive surveys, though, are expensive and not always feasible in remote areas. Known species can disappear unnoticed; others will undergo a “dark extinction”, vanishing before being scientifically recognised. The unknowns of biodiversity loss — species, timing, sites and causes — have been collectively named the “Katuš shortfall” (Katuš means to “go away” in the Yaghan language of Tierra del Fuego).
Recorded extinctions are, then, inevitably an undercount. While fewer than 1,000 of the roughly 400,000 known plant species have been declared extinct in recent centuries, according to Kew figures, the report puts 40 per cent of all plants at threat of extinction due to factors like climate change, agriculture and urban development.
Adapting techniques borrowed from animal conservation, mathematical models can calculate the probability of plant extinction using digitised databases of species names, collection dates and locations, sometimes gleaned from centuries-old travel journals. Some models now link to data on land use and habitat change. For example, a species only ever spied on a wetland site since drained for development would have a high chance of no longer existing in the wild. Antonelli explains: “Because there are so few exhaustive surveys, we now know that calculating the likelihood of species being extinct, using a number of mathematical models, gives us a much better understanding of what is happening and helps us to prioritise resources.”
Researchers in Switzerland, for example, recently revisited 197 sites across the country where the hare’s-tail cotton grass was historically recorded, finding it locally extinct in 52. By examining the characteristics of those local vanishings — such as temperature and rainfall — they predicted the extinction risk at remaining sites, which is focusing conservation efforts.
Kew’s digitised herbarium and fungarium are free to access online, in the hope that amateur naturalists and those abroad can speed up surveys, species identification and conservation. But while contributors today are motivated to save what is in danger of being lost, hoaxers have previously tried to add species that never existed.
In the 1880s, the Scots-born Augustine Henry dispatched thousands of pressed specimens to Kew during his travels in China, including a plant later found to have the flowers of one species inserted into the branch of another. Henry blamed a local collector, apologised and the new species, Actinotinus sinensis, quickly vanished for all the right reasons.