# Artificial Intelligence for Microbial Isolation and Cultivation: Progress and Challenges

## TL;DR

This paper reviews how artificial intelligence is transforming microbial isolation and cultivation methods, enabling more efficient discovery of new microbial resources.

## Contribution

The paper introduces a five-stage framework for AI's role in microbial resource discovery and highlights methodological transitions driven by AI advancements.

## Key findings

- AI technologies are enabling data-driven approaches to microbial isolation and cultivation.
- Microbial methods are shifting from passive screening to active design with AI integration.
- AI is addressing challenges like dynamic phenotypic changes and complex cultivation conditions.

## Abstract

Microbial resources are crucial for biotechnology development and fundamental scientific research. Traditional microbial techniques fail to isolate and cultivate the vast majority of microorganisms in nature, severely limiting the discovery of novel microbial resources. The rise in artificial intelligence (AI) technologies provides new computational tools to overcome bottlenecks in microbial resource discovery and utilization. This review comprehensively examines the development of AI technologies in microbial isolation and cultivation over the past three decades from the perspective of microbial resource discovery. We propose a five-stage framework: the germination period (1997–2008), the early exploration period (2008–2015), the rapid development period (2015–2019), the deep learning (DL) explosion period (2020–2022), and the AI integration period (2023–present). We focus on how AI technologies at each stage address core challenges in microbiology—including insufficient knowledge reserves, dynamic phenotypic changes, and complex cultivation conditions—through applications at the genome, individual, and community levels. Our analysis demonstrates that, as AI technologies advance iteratively, microbial isolation and cultivation methods are transitioning from experience-driven to data-driven approaches, from single-objective to systematic integration, and from passive screening to active design. This methodological transition is expanding the scope of microbial resource discovery.