# Advancing homebrew AI in diagnostic practice: opportunities and barriers†

## TL;DR

The paper discusses the potential and challenges of using locally developed AI models in pathology to improve diagnostics.

## Contribution

It highlights the need for institutional support and infrastructure to successfully implement homebrew AI in diagnostic practice.

## Key findings

- Homebrew AI in pathology offers opportunities for democratizing digital diagnostics.
- Lack of institutional backing and infrastructure may hinder the adoption of homebrew AI.
- Practical limitations and unmet needs of pathologists must be addressed for successful implementation.

## Abstract

In a recent issue of The Journal of Pathology, Calderaro et al present a timely and persuasive argument advocating for the integration of homebrew artificial intelligence (AI) models in diagnostic pathology. Their article is a robust defense of local model development within pathology departments as a pathway to democratizing digital diagnostics. This commentary expands on their premise, critically examining the real‐world implications, practical limitations, and unmet needs of practicing pathologists. The commentary outlines both the opportunities and challenges for the widespread adoption of homebrew AI in pathology practice. Without institutional backing, digital infrastructure, and sustained training efforts, the promise of homebrew AI may falter. © 2025 The Author(s). The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.