# Artificial intelligence in biomedical team science: perceptions, practices, and training needs

## TL;DR

This study explores how researchers use AI in collaborative biomedical teams and highlights the need for training and frameworks to support responsible AI use.

## Contribution

The study provides empirical insights into AI use and attitudes in biomedical team science, identifying training needs and integration challenges.

## Key findings

- AI use on research teams is highly variable, with some researchers using it daily and others never using it.
- Many researchers are concerned about misinformation, bias, and overreliance on AI tools.
- Participants expressed interest in training focused on data security, ethical use, and team integration of AI.

## Abstract

Artificial intelligence (AI) is increasingly used in biomedical research, yet limited empirical work has described how researchers use AI tools on collaborative research teams and how they view their role within team-based research. This study examines researchers’ experience with and attitudes toward AI use in collaborative research environments.

A cross-sectional survey was administered to 178 investigators engaged in collaborative research at the University of Kentucky. Questions assessed AI use across research and communication tasks, team-related decision-making practices, perceived benefits and concerns, and preferences for training and frameworks.

Thirty-nine participants responded (22%). AI use was heterogeneous: 26% had never used AI on research teams, while 42% used it weekly or daily. Nearly half reported that AI use was not discussed within teams prior to starting the work. Participants identified benefits in reducing repetitive tasks but expressed widespread concerns about misinformation, bias, and overreliance. Most participants indicated interest in self-guided training and structured frameworks, with priority topics including data security, ethical use, and practical strategies for team integration.

Findings indicate variability in both the use of AI tools on research teams and researchers’ attitudes toward their integration. Results highlight gaps between perceived benefits and current practices and suggest a need for evidence-based training and frameworks that support responsible and effective AI use on collaborative research teams.