# Artificial Intelligence in Food Safety: A Tertiary Study

## TL;DR

This paper reviews how artificial intelligence is being used to improve food safety, highlighting trends and future opportunities.

## Contribution

The study is the first tertiary analysis synthesizing insights from secondary studies on AI applications in food safety.

## Key findings

- Dairy products are the most studied food category in AI-based food safety research.
- Neural networks are the most commonly used AI algorithms in this domain.
- Most AI applications focus on detecting chemical hazards rather than biological or physical ones.

## Abstract

Food safety remains a critical factor in preventing contaminated and hazardous products from reaching consumers. The integration of artificial intelligence (AI) and its capacity to deal with vast datasets has significantly enhanced food safety protocols, and a substantial number of primary and secondary studies have emerged at the intersection of these two domains. Although several studies have addressed AI applications in food safety, no tertiary study has yet synthesized the collective insights from existing systematic reviews. To address this gap, this paper provides a comprehensive overview of the current state of AI applications in food safety through a systematic tertiary analysis of secondary studies. By systematically analyzing secondary studies, this research identifies key trends such as the food categories most frequently investigated, the data sources utilized, prevalent food safety hazards, the commonly adopted AI algorithms, and the challenges associated with their implementation within the field. The analysis revealed that dairy products received the greatest research attention, with sensing data serving as the primary data source. Neural networks emerged as the predominant AI approach. Furthermore, most applications focused on the detection of chemical food safety hazards rather than biological, physical, or general predictive modeling. Notably, this study highlights a lack of AI algorithms utilizing unstructured data, despite its growing relevance in the era of generative AI. Accordingly, future research directions are discussed, particularly the transformative potential of large language models (LLMs) in food safety monitoring and regulatory compliance.