Determining whether a plant or fungus has gone extinct remains a complex challenge for scientists, who must contend with factors such as elusive growth habits, changes in habitat, and incomplete survey data. This issue is at the center of a recent report published by the Royal Botanic Gardens, Kew, which highlights how advances in digitisation and artificial intelligence (AI) are reshaping conservation efforts and revealing that the loss of species may be significantly underestimated.

Alexander Antonelli, executive director of science at Kew, explained at the launch of the 2026 State of the World’s Plants and Fungi report that definitively proving a species no longer exists is difficult. Instead of relying solely on binary declarations of extinction, researchers are increasingly employing mathematical models that estimate the likelihood a species persists. The digitisation of Kew’s extensive collection—comprising 7.4 million specimens of plants and fungi—combined with AI tools, allows scientists to link specimens to extensive searchable data relevant to assessing species survival and extinction risk.

The report emphasizes the importance of protecting biodiversity, which underpins ecosystems vital to human welfare, including water supply, food security, and the provision of natural resources. Beyond ecological and economic considerations, preserving threatened plants and fungi may lead to the discovery of valuable compounds, such as new medicines or substances capable of breaking down plastics.

The International Union for Conservation of Nature (IUCN) defines a species as extinct in the wild only after exhaustive surveys conducted across its known range fail to find any individuals. However, such comprehensive surveys are costly and logistically challenging, especially in remote or inaccessible areas. Some species may vanish unnoticed before scientists ever document them, a phenomenon termed “dark extinction.” Collectively, these unknowns regarding biodiversity loss are referred to as the “Katutis shortfall,” named after a Yaghan word meaning “go away.”

Recorded extinctions represent only a fraction of actual losses. While fewer than 1,000 of the roughly 400,000 known plant species have been declared extinct in recent centuries, Kew’s report estimates that approximately 40% of all plants are currently at risk, driven by climate change, agricultural expansion, and urban development.

Drawing on methodologies from animal conservation, researchers use digitised historical records, including species names, collection dates, and locations—sometimes extracted from centuries-old travel journals—to feed extinction probability models. These models increasingly incorporate data on habitat changes, such as land use and environmental shifts. For instance, a species previously documented solely in a wetland now drained for development would be assigned a high extinction probability.

In Switzerland, researchers recently revisited 197 historical sites of the hare’s-tail cotton grass, finding it locally extinct in 52 locations. By analyzing environmental factors influencing these local disappearances, scientists developed predictive models to prioritize conservation efforts where the species still exists.

Kew’s herbarium and fungarium collections are freely accessible online, aiming to engage amateur naturalists and international collaborators in accelerating species surveys, identification, and protection. However, safeguarding data accuracy remains important, as historical instances of hoaxes have occurred. One notable case from the 1880s involved Scots-born botanist Augustine Henry, whose specimens from China included a plant later revealed to be a fabricated hybrid formed by combining flowers of one species with the branches of another—an act attributed to a local collector and since corrected.

As tools such as digitisation and AI evolve, scientists are gaining improved capacity to understand extinction risks and mobilize conservation resources more effectively, addressing the complex and urgent task of preserving global plant and fungal biodiversity.