Fwd: AI Meets Biodiversity: iChatBio Webinar

[Katelin Pearson]

---------- Forwarded message ---------
From: Davison, Shyla <sdav...@floridamuseum.ufl.edu>
Date: Wed, Aug 13, 2025 at 9:49 AM
Subject: AI Meets Biodiversity: iChatBio Webinar
To: IDIG...@LISTS.UFL.EDU <IDIG...@lists.ufl.edu>

Hello!

Curious how AI is transforming biodiversity research? Join us for a look at iChatBio, a cutting-edge agentic system designed to help scientists discover, process, and analyze biodiversity data with ease. Explore how this innovative system streamlines tasks such as species occurrence summaries, dataset refinement, and data retrieval from iDigBio, enhancing both productivity and reproducibility in science.

Don’t miss this exciting peek into the future of research!

Register here: https://ufl.zoom.us/webinar/register/WN_eeerYXPRSWObZZiXA6OZdQ

_________________________________________________________________________________________________________________

Abstract: iChatBio: A Modular Agent System for Biodiversity Data Exploration 

As large language models (LLMs) gain new capabilities through function calling and tool use, a new class of systems—agentic systems—has emerged. These systems combine natural language with structured reasoning and API execution, enabling AI agents to fulfill user requests by orchestrating tools, fetching data, and returning results in human-readable form. As research tools, agentic systems have the potential to transform how researchers interact with scientific data. Researchers can express what information they need (e.g., “What collections contain specimens of Danaus Plexippus?”) without specifying where or how to find it, while specialized agents handle the labor-intensive task decomposition, data retrieval, and processing across heterogeneous sources.

General-purpose agentic systems lack the domain expertise required to make effective use of scientific data. This presentation highlights recent developments in iChatBio, an agentic system designed to assist biodiversity researchers with discovering, processing, and analyzing biodiversity-related information. Unlike general-purpose tools, our approach emphasizes transparency, modularity, and domain-specific validation to ensure that results are trustworthy and citable. We demonstrate how targeted agents can streamline tasks such as data retrieval from iDigBio, species occurrence summarization, and dataset refinement, advancing both usability and reproducibility in biodiversity science.

Thanks,
Shyla

		Shyla R. Davison, Education and Outreach Coordinator Integrated Digitized Biocollections (iDigBio) University of Florida Florida Museum of Natural History Sdav...@ufl.edu

---

To unsubscribe from the IDIGBIO-L list, click the following link:
https://lists.ufl.edu/cgi-bin/wa?SUBED1=IDIGBIO-L&A=1

--

Katie D. Pearson

Project & Data Manager

Symbiota Support Hub, iDigBio