Unlocking biodiversity data with robotic imaging and AI-driven transcription of natural history collections.

Authors

Christine A Johnson, Erin M Willigan, Ellen Y Hwang, Parker M Austin, Kyle L Chang, Peter Daly, Ann Davis, Jovanni M Gonzalez, Chia-Ni Kao, Isobel E J Mifsud, Kitri A Miller, Marinela Nicolescu, Ana Maria Ruiz, Elliott Y Ruo, Julia L Schwartzman, Alyssa B Seeman, Janice Stenzel, Natalia C Zurita, Lily F Berniker, Estefanía Rodríguez

Published in

Science advances. Volume 12. Issue 18. Pages eaeb6975. Epub Apr 29, 2026.

Abstract

Digitizing metadata on natural history specimen labels remains a critical bottleneck for biodiversity research. We present a transformative workflow integrating robotic imaging with artificial intelligence (AI)-driven transcription for rapid, comprehensive data extraction from specimen labels. Single high-resolution images of specimens and associated labels were submitted to Gemini 2.5 Flash and GPT-4 Turbo to extract verbatim textual information. This approach yielded ~600 verbatim transcriptions per hour, a 30-fold increase in efficiency compared to traditional manual methods, which yielded ~20 transcriptions per hour. Releasing historical specimen metadata facilitates information accessibility and provides temporal and spatial context for a variety of analyses. Our method fosters the reconnection of disparate biological datasets previously segregated among departments or institutions to unite ecologically interdependent components (e.g., host/parasite and pollinator/plant) for a more complete understanding of biodiversity dynamics.

PMID:
42054451
Bibliographic data and abstract were imported from PubMed on 30 Apr 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.