In the herbariums of Kew, which house millions of dried specimens including those collected by Darwin himself, what is emerging is more than just an archive; it is a living map of potential salvation. Old sheets of yellowed plants that gathered dust in cabinets for decades are now being scanned and uploaded to databases, accessible to any researcher with a computer.
For centuries, botanical collections remained inaccessible to the general public, with millions of specimens seen only by specialists. Digitization and artificial intelligence are changing the rules of the game. AI models can already identify species of sedge and peat moss—where differences are visible only under a microscope—and are finding vulnerable or undescribed plants faster than ever.
According to a report by the Royal Botanic Gardens, Kew, approximately forty percent of the seventy thousand assessed plant species are threatened with extinction. Another three hundred thirty thousand species have not yet been evaluated, while experts estimate that roughly one hundred thousand plants remain without a scientific name. For fungi, the picture is even blurrier: only a fraction of the estimated two million species has been described, and the extinction risk has been assessed for less than one percent of known varieties.
Kew has fully digitized its seven and a half million specimens, including Darwin’s own collections. At the height of the project, the team captured up to twenty thousand high-resolution images per day. While one hundred forty-five million digital specimens are now available worldwide, this represents less than sixteen percent of all materials stored in herbariums globally.
Digitized collections have already revealed how the rhythm of plant life is shifting. A model that analyzed eight million specimens found that over the last century, flowering times have shifted by an average of two and a half days per decade. As some species bloom earlier and others later, these changes can disrupt critical links with pollinators and animals that depend on seasonal cycles.
Modern technology allows researchers to extract DNA even from fungal specimens up to one hundred eighty years old. Such data aids in the search for new medicinal compounds and provides a better understanding of how fungal diseases spread. AI accelerates work with difficult plant groups and provides access to materials from high-biodiversity regions like Madagascar.
The report's authors also highlight certain risks, such as the high energy and water consumption of data centers. Without digital tools, however, a significant portion of these collections will remain invisible to science. Furthermore, digital databases could exacerbate existing biases if they remain incomplete.
For these systems to be truly effective, more comprehensive collections, partnerships between tech companies and conservation organizations, and sustainable funding for archives are essential. The digitization of Darwin’s collection is just the beginning of a journey where ancient specimens may point the way toward plants capable of fighting hunger and disease.
Every digitized page is more than just a photograph; it is a potential key to species that might vanish before they can even be studied.
