Damon Little, Ph.D., is the Curator of Bioinformatics in the Center for Biodiversity & Evolution at the New York Botanical Garden.

---

Vast numbers of plant and fungal species are still unknown to science, while biodiversity loss and climate change accelerate. Royal Botanic Garden, Kew’s sixth State of the World’s Plants and Fungi report explores how digitized collections and new technologies are helping scientists speed up the discovery and assessment of species before they are lost. NYBG has been integrating new technologies and AI into advancing botanical research for years and is proud to contribute a number of papers to this pursuit. 

AI holds great promise for accelerating the rate of botanical discovery. AI stands to not only automate manual and repetitive tasks, but also to facilitate new avenues of research. The first thing that comes to mind, when most people think of AI, is a Large Language Model (LLM), produced by a big tech company, that can (attempt) to answer any question put to it. Although these models are often very good at general tasks and contain a startling amount of specialized knowledge, they have not been specifically designed to support scientific enquiry. Small AI models designed for scientific research often perform better than their large general counterparts for these specific scientific tasks.  

Scientists at NYBG have been hard at work building responsible specialized models to support botanical research. Some of these specialized models have been designed to automate routine scientific tasks such as sorting herbarium specimen images into categories so that further analyses use directly comparable specimens, or isolating specific parts of micro-CT images for study. Other AI models have been used to augment human abilities—to help scientists see the unseen, much like a microscope magnifies what the human eye cannot directly see—and find previously unrecognized patterns in specimen images. For instance, peat mosses (Sphagnum) usually require dissection and study under a microscope to determine which species they are, but NYBG scientists have trained an AI model to recognize peat moss species without dissection or magnification. 

The key to building high-performing AI models is carefully curated digital data—something NYBG scientists abundantly produce in the Library, Herbarium, and Laboratory. Although these data are often gathered with a specific scientific question in mind, in many instances they can later be repurposed and combined with data gathered for other research projects to answer scientific questions that the original researchers did not imagine. As a result of this synergy, open sharing of digital data is increasingly important for all types of botanical research. 

Botanists everywhere were excited to learn that the Royal Botanic Gardens, Kew, which houses the largest preserved collection of plants and fungi in the world, was planning to digitize their collections and make them freely available. To celebrate the complete digitization, more than 400 experts from 40 countries collaborated to produce a report on the State of the World’s Plants and Fungi. In addition, a special collection of scientific papers focusing on the benefits of digitized collections was published in two influential scientific journals: New Phytologist and Plants, People, Planet. NYBG scientists contributed to both endeavors. 

Papers in the collection: 

Antonelli, A., C. Davis, I. Larridon, D.P. Little, R.J. Smith, S. Smith. 2026. “The digital biodiversity revolution.” New Phytologist. 

Arno, J., J. Morel, F. Rasaminirina, J. de Fátima Maciel-Silva, D. Cahen, D.P. Little, D. Silvestro, A. Antonelli, O. Grace, L. Zhang, and I. Larridon. 2026. “A pipeline to compile expert-verified datasets of digitised herbarium specimens for automated plant identification to accelerate taxonomy.” Plants, People, Planet. 

Little, D.P., B. Aguero, A.J. Shaw, and M. Tessler. 2026. “AI for difficult herbarium specimens: identification of peat mosses (subgenus Sphagnum) without dissection.” New Phytologist. 

Ávila, F.A., J. Park, L. Feder, D.P. Little. 2026. “Herbariograph: a deep-learning tool to classify specimen images.” New Phytologist. 

Tessler, M. and D.P. Little. 2026. “On herbarium specimen images and artificial intelligence.” New Phytologist.