Artificial intelligence has helped reveal how plants are shifting flowering times across the world due to climate change, a new report has said. A global study using AI to analyse eight million digitalised plant specimens dating back a century revealed flowering has shifted by 2.5 days earlier or later per decade on average, disrupting relationships between plants and pollinators.
It is just one of the ways in which converting preserved specimens of plants and fungi that lie 'hidden in cupboards and boxes' into digital records is transforming the fight to save life on Earth, experts at the Royal Botanic Gardens, Kew, have said. New tech is being applied to collections of plants and fungi, from digitising specimens to speed up the search for climate-resilient wild relatives of key food plants such as coffee, to genome sequencing long-preserved fungal species to find new medicines or alternatives to meat, the experts said.
The approaches are highlighted in Kew science team's latest 'state of the world's plants and fungi' report, marking the 10th anniversary since the first such report, and drawing on work by 400 scientists in 40 countries. The report warns the threats to plant and fungi species are significantly underestimated, with less than a fifth of plants and only a tiny fraction – just 0.6% – of fungi assessed for their extinction risk.
An estimated 100,000 plants and two million fungi species are still to be discovered, with 4,600 new species of plant and 7,800 new fungi described by scientists for the first time in 2024 and 2025. Kew's executive director of science Alexandre Antonelli said the new discoveries were 'just scratching the surface' and many of the newly described species are threatened with extinction 'from day one'.
But Kew's sixth state of the world's plants and fungi report has a hopeful message, focusing on the power of AI, digitisation and other technology to boost knowledge, information sharing and conservation efforts around the world – and its potential to do more with coordinated international action. It comes as Kew completes its project to digitise of its herbarium and fungarium, using high resolution photography to turn 7.4 million specimens of pressed plant leaves, flowers, seedheads, mushrooms and spores dating back centuries into digital records that can be freely accessed and searched online.
The scheme, funded by the Environment Department (Defra), opened every cupboard and box in the herbarium to scan the specimen sheets at 40 imaging stations, digitally recording up to 20,000 images a day, while bigger species such as palms had to removed from their boxes and carefully recorded. Kew's four-year project has thrown up insights of the past in its collections, from specimens collected and logged by famed naturalist Charles Darwin to the 6,700 specimens gathered by First World War service personnel.
The digital specimens now form part of a network of 145 million records from institutions and organisations worldwide that is allowing researchers across the world to access important collections that once they would have had to travel thousands of miles to see. Professor Antonelli said: 'We can use digital assets, artificial intelligence and other technologies to really harness the information locked in many of these specimens that have been here for centuries, and use that to advance science and conservation at a global level.'
He added: 'We can use this digitised information to discover new species, and also to reveal species that have gone extinct or are likely to have done so. We can track change in relation to climate change and we can also unlock the information in those collections in a way that is much more equitable and accessible to anyone, anywhere for free.' In Costa Rica, researchers increased the country's known fungal diversity by 20% by combining published records with digital collections, while AI can be trained to distinguish plants such as sedges and mosses to help scientists identify species more quickly.
The first global study of the impact of climate change on flowering times trained an AI model to classify images for the presence or absence of flowers and then applied it to eight million specimens of 200,000 species. It revealed not only a 2.5 day average shift earlier or later per decade over the course of a century but also variations between regions, with greater changes in the tropics. Digitised records have also boosted work on changes to flowering and fruiting of plants in the tropics and the Arctic where fieldwork is difficult, been used to map loss of species, safeguard protected areas and help with food security, the experts said.
Prof Antonelli pointed to research previously carried out by Kew to track down climate-resilient alternatives to the current commercial coffee species. That research was done using the specimens in Kew's herbarium and took a long time, but with digitisation 'we could do it much faster with other species,' he said. Scientists are also using digitised herbarium records, sighting histories and statistical models to estimate whether species are extinct or simply undetected. And for the first time scientists are unlocking the 'dark matter of fungi', producing high quality genomes from very old fungal specimens in Kew's fungarium, including samples up to 180 years old.
The experts said the ability to unlock the secrets of known species of fungi – which make up less than 10% of the estimated species out there – could lead to multiple new uses from high-protein alternatives to meat to protecting crops from pathogens and finding plastic-digesting fungi to tackle pollution.
