Introduction

Foodborne outbreak investigations and surveillance used to rely on conventional methodologies such as manual data collection, laboratory testing, and epidemiological analysis. These approaches are often time-consuming, resource-intensive, and limited by the quality and timeliness of data.

Recent advancements in technology, such as machine learning, natural language processing, and whole genome sequencing, offer innovative solutions to overcome these challenges and enhance food safety [1]. In 2021, there was a notable increase in reported cases of zoonotic diseases and foodborne outbreaks compared to 2020, although the overall number of cases remained lower than pre-pandemic figures [2]. Given the risk of rising cases, timely surveillance and effective outbreak investigation remain critical to the prevention and control of foodborne diseases [2]. These activities are primarily conducted by public health agencies staffed with professionals skilled in epidemiology, food safety, nutrition, and laboratory sciences. Collaboration across subnational, national, and international levels enhances the effectiveness of these efforts [3]. Prior to the COVID-19 pandemic, foodborne outbreak investigations and surveillance largely relied on conventional methodologies (e.g., manual data collection, laboratory testing, and epidemiological analysis), which were often time-consuming, not always cost-effective, and sometimes limited by the scale and speed of data processing.

Today, the realm of foodborne disease surveillance and outbreak investigations is evolving in complexity. This includes the disciplines of genomics and bioinformatics, especially after the adoption of advanced technologies such as whole genome sequencing [1]. Also, there is a shift in food production, distribution, and consumption patterns, which demand a multidisciplinary approach, which is driven by a complex number of factors that are sometimes interconnected, such as urbanization, environmental factors, climate change [4], and the scarcity of human resources. Therefore, it is critical to provide adequate training and support to the workforce for effective surveillance and management of outbreaks [3], even in high-income countries.

Furthermore, prompt detection and intervention in cases of foodborne disease outbreaks are vital not only for stopping ongoing outbreaks but also for preventing potential future outbreaks, thereby decreasing the overall occurrence of foodborne diseases. In this paper, we discuss the potential role of artificial intelligence (AI) in enhancing four key areas of foodborne disease prevention and management: predictive analytics, supply chain monitoring, public health surveillance, and laboratory investigations.

Methods

We conducted a mini review of the literature to explore the role of AI in the context of foodborne diseases. The literature search was performed using PubMed and Google Scholar. We used a combination of keywords and Medical Subject Headings (MeSH) terms related to two main concepts: “foodborne diseases” and “artificial intelligence”. Articles were selected based on their relevance to AI applications in the prevention, detection, and management of foodborne disease outbreaks, including outbreak investigations and food safety. We included articles published in English and applied no exclusions for the date of publication. No time scope was used.

The four emerging themes included: 1) general predictive models; 2) the introduction of AI in the food supply chain; 3) the role of AI in enhancing public health surveillance; and 4) AI can be used for foodborne outbreak laboratory investigations (Table 1).

Challenges

The utility of AI, e.g., machine learning, has its own limitations as well. We list them under the following categories (Table 2):

• 1)

  Data integrity and model performance challenges, which are related to the quality, reliability, and scalability of data:

  • i)

    Data quality is crucial in food safety, because it affects each of the performance and accuracy of AI and machine learning models;

  • ii)

    The need for data validation, by testing AI models on external datasets [24];

  • iii)

    The scalability of AI solutions to manage large databases and adaptation to changing conditions can be a challenge. The needs of food safety evolve as the volume of data increases; therefore, AI needs to be agile to meet the needs of different scenarios of new types of data [25];

  • iv)

    Machine learning, as a subset of AI, learns from experiences [26] and preexisting data to train its models, so it can be used for new data [26], which implies that data quality is crucial [27];

  • v)

    Bias in AI technology can lead to unfair or unjust outcomes. For example, biased data used in AI models’ training can have inaccurate predictions. This can affect certain groups in an unjust fashion, including decision-making on product quality and evaluations of suppliers [28].

• 2)

  System integration and operational feasibility, which include the logistical/practical impediments in deploying AI into the current food system, including legacy infrastructure, resources, and compliance:

  • i)

    Integrating AI with existing food systems can be complex due to incompatibility with the traditional systems [28];

  • ii)

    Adoption of AI-related technologies, due to either high costs or complex regulatory frameworks and standards [29];

  • iii)

    The utility and deployment of AI, including the technologies needed for data collection, can have an effect on the environment, including energy consumption and electronic waste.

• 3)

  Ethics, transparency, and trust, which include safeguarding rights and ensuring the ethical deployment of AI in food safety contexts:

  • i)

    The AI algorithms tend to be perceived as “black boxes”, which means the decision-making process is not fully interpretable. Lack of transparency makes it impossible to understand the rationale behind AI decisions. The transparency is crucial to increase trust and compliance with processes [30];

  • ii)

    Ethical considerations, such as data privacy and informed consent, are essential for accountability [31]. The use of AI includes the collection of data from various sources (e.g., consumer feedback or health-related information) in the food safety process. Mismanagement or unauthorized access of this data is a potential risk for privacy violations, which can lead to the loss of consumer trust [25];

  • iii)

    Informed consent is required before the collection of personal data. This is a crucial ethical principle to be considered [26].

Conclusions

In this review, we summarized the enablers and impediments of using AI in food safety (Table 3).

AI has the potential to transform real-time surveillance and the automation of complex tasks; yet, the success of its implementation depends on high-quality datasets, transparent algorithm design, and scalable infrastructure.

There are some potential challenges to utilizing AI in public health, as current algorithms may not be sufficiently advanced to ensure the resolution of complex tasks. However, initiating the shift from traditional, manual methods to AI-driven approaches will provide an efficient and more accurate potential for safeguarding public health. Advancements in digital data management, alongside innovative solutions for data privacy, are paving the way for the increased adoption of AI. With continued digital upskilling of the staff and strategic industry collaborations, AI is set to become a crucial decision-support asset in food safety, complementing human expertise.

Finally, investing in training staff on AI and ethical governance is considered a strategic investment to ensure that AI becomes a trusted, equitable, and sustainable decision-support tool that complements and enhances public systems.