# Artificial intelligence in functional food innovation: Bioactive enhancement and formulation optimization: A quasi-systematic review

## TL;DR

This paper reviews how AI is used in functional food research, focusing on bioactive compound prediction and formulation optimization.

## Contribution

It highlights the novel use of explainable AI and deep learning in enhancing functional food precision and antioxidant development.

## Key findings

- Explainable AI improves transparency in predicting bioactive compounds.
- Deep learning and omics integration enhance functional food precision.
- Convolutional neural networks are effective for analyzing complex metabolomic data.

## Abstract

Artificial intelligence (AI) is increasingly integrated into functional food research. This quasi-systematic review analyzes 53 peer-reviewed studies (2015–2025) to outline current applications and emerging directions, including the underexplored domain of antioxidant food development. The review attempts to provide an updated synthesis of AI approaches across compound discovery, metabolomics, and consumer modeling, emphasizing knowledge gaps and opportunities for methodological integration. Data-driven AI (classical machine learning) and deep learning methods have been applied to predict antioxidant activity, identify bioactive compounds, and reveal patterns in metabolomic data. Unsupervised approaches have assisted in clustering complex datasets, whereas optimization algorithms supported the adjustment of sensory, nutritional, and functional attributes. However, many current systems remain limited to in silico findings, lacking experimental or clinical validation. Consumer modeling remains largely predictive, with limited integration of ethical and regulatory dimensions. Continued collaboration between food scientists and data scientists is essential for translating computational insights into practical applications.

•A quasi-systematic review of 53 studies on AI in functional food development.•Explainable AI ensures transparency in bioactive compound prediction.•Deep learning and omics integration enhance functional food precision.•Convolutional neural networks excel in analyzing complex metabolomic data.•Microbiota variability challenges AI-based personalized nutrition models.

A quasi-systematic review of 53 studies on AI in functional food development.

Explainable AI ensures transparency in bioactive compound prediction.

Deep learning and omics integration enhance functional food precision.

Convolutional neural networks excel in analyzing complex metabolomic data.

Microbiota variability challenges AI-based personalized nutrition models.