Policy Framework and Implementation Guidelines for Agentic GenAI Integration in Food Safety Systems

*Corresponding author: Satyadhar Joshi Copyright © 2025 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0. Policy Framework and Implementation Guidelines for Agentic GenAI Integration in Food Safety Systems Satyadhar Joshi * Independent Researcher, Alumnus I-MBA, Bar Ilan University, Israel, Alumnus MS in IT Touro College NYC, USA. World Journal of Biology Pharmacy and Health Sciences, 2025, 23(03), 303-315 Publication history: Received on 10 August 2025; revised on 15 September 2025; accepted on 18 September 2025 Article DOI: https://doi.org/10.30574/wjbphs.2025.23.3.0845 Abstract This paper presents a comprehensive review of artificial intelligence (AI) applications in food safety and quality control, focusing on emerging technologies including generative AI, agentic AI systems, and automated compliance solutions. This review synthesizes current research and industry applications, highlighting how AI-driven systems are transforming food safety protocols, enhancing regulatory compliance, and improving overall food quality management focusing on last two years. We examine various AI implementations, from optical imaging for bacterial detection to intelligent compliance agents and generative AI for supply chain optimization. This paper synthesizes current research and industry applications across multiple domains: automated visual inspection systems that detect contaminants with precision exceeding human capabilities; predictive quality analytics that forecast potential safety issues before they manifest; AI-driven regulatory compliance systems that continuously monitor and interpret complex regulatory requirements; and autonomous agentic systems that make real-time decisions without human intervention. The review also addresses significant technological innovations, including the FDA’s development of AI tools for regulatory operations, generative AI applications for scenario planning and documentation, and cloud-based AI architectures deployed across major platforms. Critical challenges are examined, including data quality requirements, regulatory validation frameworks, system integration complexities, and ethical considerations. The paper concludes with policy recommendations for government implementation, proposing structured approaches to AI validation, data sharing incentives, regulatory modernization, research support, and ethical oversight. Keywords: Artificial Intelligence; Food Safety; Quality Control; Generative AI; Agentic AI; Autonomous Systems; Predictive Analytics; Regulatory Compliance; Policy Framework 1. Introduction Traditional methodologies for ensuring safety—relying on manual inspections, periodic laboratory testing, and reactive protocols—are increasingly proving to be inadequate. These methods are often time-consuming, labor-intensive, susceptible to human error, and struggle to keep pace with the complexity of modern global supply chains. The consequences of these limitations can be severe, ranging from costly product recalls and brand damage to widespread foodborne illness outbreaks. The integration of machine learning (ML), computer vision (CV), and the Internet of Things (IoT) has already begun revolutionizing this field by enabling unprecedented capabilities in real-time monitoring, predictive analytics, and automated anomaly detection [1] and [2]. These technologies facilitate the continuous analysis of vast datasets from sensors and imaging systems, identifying potential contamination and quality issues far more rapidly and accurately than previously possible [3], [4]. World Journal of Biology Pharmacy and Health Sciences, 2025, 23(03), 303-315 304 While significant, these advancements represent only the initial wave of AI’s transformative potential. This paper argues that the next frontier lies in the adoption of two groundbreaking technologies: Generative AI and Agentic AI systems. Generative AI moves beyond analysis to creation, capable of simulating novel contamination scenarios, generating optimized supply chain models, and automating complex compliance documentation [5], [6], [7]. Concurrently, Agentic AI systems are evolving from passive tools into autonomous actors that can perceive their environment, make decisions, and execute actions—from automatically adjusting processing parameters to managing real-time compliance protocols—with minimal human intervention [8], [9], [10]. The recent initiative by the U.S. FDA to launch an agencywide AI tool underscores the regulatory recognition of this transformative potential [11]. 2. Background on AI Technologies in Food Safety Artificial intelligence encompasses a broad range of technologies that enable machines to perform tasks that typically require human intelligence. In the context of food safety and quality control, several AI technologies have proven particularly valuable. 2.1. Machine Learning and Predictive Analytics Machine learning algorithms can analyze vast amounts of data from various sources, including sensor networks, production records, and environmental monitoring systems, to identify patterns and predict potential safety issues [2]. These predictive capabilities enable proactive interventions before problems escalate into full-blown safety incidents. 2.2. Computer Vision and Optical Imaging Advanced computer vision systems, often combined with optical imaging technologies, can detect contaminants, identify defects, and monitor food quality parameters with exceptional accuracy and speed [3]. These systems can identify bacterial contamination, physical defects, and quality issues that might be invisible to the human eye. 2.3. Natural Language Processing for Compliance Natural language processing (NLP) technologies enable AI systems to understand, interpret, and process regulatory documents, compliance requirements, and safety protocols [12]. This capability is particularly valuable for managing the complex and constantly evolving landscape of food safety regulations. 2.4. Internet of Things (IoT) Integration AI systems integrated with IoT sensors can monitor environmental conditions, equipment performance, and product quality in real-time throughout the food supply chain. This continuous monitoring provides unprecedented visibility into food safety parameters. The convergence of these technologies creates powerful systems capable of transforming traditional food safety practices from reactive to proactive, from manual to automated, and from periodic to continuous monitoring. 3. Discussion: Architectural Analysis of AI Systems The integration of AI into food safety is not merely a collection of isolated tools but requires a cohesive architectural framework. The proposed figures in this review illustrate the evolution from conceptual models to practical, cloudbased implementations of these intelligent systems. World Journal of Biology Pharmacy and Health Sciences, 2025, 23(03), 303-315 305 Figure 1 Comprehensive Architecture of an AI Agent System for Food Safety and Quality Control. The system integrates diverse data inputs (IoT, Vision, Supply Chain, Regulatory) which are preprocessed and analyzed by a Generative AI core The AI Agent Controller uses these insights, along with a dedicated Knowledge Base and predictive engines, to make autonomous decisions. These decisions are executed through an action framework, generating alerts, reports, and process adjustments. Crucially, dashed feedback loops show how outcome data continuously improves the system's predictive models and knowledge. Figure 1: provides a foundational overview of a single AI agent’s architecture, highlighting the critical data flow from diverse inputs (IoT sensors, vision systems, regulatory feeds) through a generative AI core for pattern recognition, culminating in autonomous decision-making and action execution. This blueprint emphasizes the closed-loop feedback mechanism essential for continuous learning and improvement, a cornerstone of effective agentic systems. Building upon this core concept, Figure 2 presents a comprehensive, multi-layered architecture for an enterprise-wide AI-powered quality control system. It effectively scales the agentic paradigm by delineating four distinct layers: Data Input, AI Processing, Agentic Systems, and Actions/Outputs. This holistic view demonstrates how specialized agents (e.g., Quality Control, Food Safety, Compliance) operate in tandem, each powered by dedicated AI components (e.g., Predictive Analytics, Knowledge Graphs), to manage the entire spectrum of food safety operations. The architecture underscores the necessity of a modular yet integrated approach to handle the complexity of modern food production. Finally, Figure 3 translates these architectural principles into a practical, scalable deployment model using cloud infrastructure. It demonstrates how the agentic functions can be distributed across major cloud service providers (AWS, Azure, Google Cloud), highlighting the industry trend towards leveraging scalable computing resources and managed AI services. This model is crucial for understanding how organizations, especially those without extensive in-house IT infrastructure, can adopt and benefit from these advanced technologies. The cloud-based approach facilitates scalability, reliability, and easier integration of the continuous feedback loops that are vital for system evolution. 4. AI Applications in Quality Control and Inspection 4.1. Automated Visual Inspection Systems AI-powered visual inspection systems have revolutionized quality control in food processing facilities. These systems use advanced computer vision algorithms to detect defects, contaminants, and quality issues with precision exceeding human capabilities [13]. For instance, optical imaging combined with AI can rapidly identify bacterial contamination in food products, significantly reducing the risk of foodborne illnesses [3]. World Journal of Biology Pharmacy and Health Sciences, 2025, 23(03), 303-315 306 Figure 2 Architecture of an AI Agent System with Generative AI Core for Food Safety and Quality Control. The system integrates multiple data sources, processes information through generative AI and autonomous decision engines, and executes various food safety actions with continuous feedback loops for learning and improvement The implementation of these systems has shown remarkable results in various food sectors. In produce processing, AI systems can sort fruits and vegetables based on quality parameters, while in meat processing, they can detect abnormalities and contaminants that might be missed by human inspectors [14]. The speed and accuracy of these systems not only improve food safety but also enhance operational efficiency. 4.2. Predictive Quality Analytics AI systems can predict quality issues before they manifest in finished products. By analyzing data from multiple sources, including raw material quality, processing parameters, and environmental conditions, machine learning algorithms can identify patterns that precede quality deterioration [2]. This predictive capability enables proactive adjustments to processing parameters, reducing waste and ensuring consistent product quality. Predictive analytics also play a crucial role in shelf-life estimation and freshness management. AI algorithms can analyze various factors affecting product shelf life and provide accurate predictions, enabling better inventory management and reducing food waste [15]. 4.3. Real-time Process Monitoring and Control AI systems enable real-time monitoring and control of food processing operations. These systems can continuously analyze process parameters and make automatic adjustments to maintain optimal quality conditions [16]. For example, in thermal processing operations, AI can dynamically adjust temperatures and processing times based on real-time quality measurements, ensuring both safety and quality objectives are met. The integration of AI with process control systems represents a significant advancement over traditional statistical process control methods. AI systems can handle complex, non-linear relationships between process parameters and quality outcomes, leading to more precise control and consistent quality [10]. World Journal of Biology Pharmacy and Health Sciences, 2025, 23(03), 303-315 307 5. AI-Driven Compliance and Regulatory Systems 5.1. Automated Regulatory Compliance The complex and evolving nature of food safety regulations presents significant challenges for food manufacturers. AI systems are increasingly being deployed to manage regulatory compliance by continuously monitoring regulatory updates, interpreting requirements, and ensuring that operational practices remain compliant [12]. These systems can analyze thousands of regulatory documents from multiple jurisdictions, identify relevant requirements, and translate them into actionable compliance tasks. AI-powered compliance systems significantly reduce the administrative burden associated with regulatory compliance while improving accuracy and completeness. They can automatically generate compliance reports, maintain audit trails, and provide real-time compliance status updates [17]. This automation not only reduces costs but also minimizes the risk of compliance failures that could lead to recalls or regulatory actions. 5.2. HACCP Plan Development and Management AI technologies are transforming Hazard Analysis and Critical Control Points (HACCP) plan development and implementation. AI systems can analyze historical data, scientific literature, and regulatory requirements to identify potential hazards and recommend appropriate control measures [18]. These systems can also monitor critical control points in real-time, automatically triggering corrective actions when deviations occur. The use of AI in HACCP management enhances the scientific basis of food safety plans while making them more dynamic and responsive to changing conditions. AI systems can continuously learn from new data, improving their hazard identification and control recommendation capabilities over time [19]. 5.3. Traceability and Recall Management AI-enhanced traceability systems provide unprecedented capabilities for tracking food products throughout the supply chain. These systems can quickly trace the movement of products from farm to fork, enabling rapid response to safety incidents and efficient management of product recalls [20]. When a safety issue is identified, AI systems can instantly identify affected products, their locations, and distribution patterns, significantly reducing the time and scope of recalls. Blockchain technology integrated with AI creates immutable audit trails and enhances the reliability of traceability data. This combination provides transparent and verifiable records of food safety practices throughout the supply chain [21]. 6. Emerging Technologies: Generative AI and Agentic Systems 6.1. Generative AI in Food Safety Generative AI represents a groundbreaking advancement in artificial intelligence, with significant implications for food safety and quality control. Unlike traditional AI systems that recognize patterns in existing data, generative AI can create new content, simulate scenarios, and generate innovative solutions to complex problems [5]. In the context of food safety, generative AI applications include: • Predictive scenario modeling: Generating potential contamination scenarios and their outcomes to develop more robust prevention strategies [22] • Automated documentation generation: Creating compliance documents, safety protocols, and training materials tailored to specific operations and regulatory requirements [6] • Supply chain optimization: Generating optimal supply chain configurations that minimize safety risks while maintaining efficiency [7] Generative AI also enables the creation of synthetic data for training purposes, addressing the challenge of limited realworld data for rare safety events. This capability enhances the training of other AI systems without compromising real operational data [23]. World Journal of Biology Pharmacy and Health Sciences, 2025, 23(03), 303-315 308 6.2. Agentic AI Systems Agentic AI represents a paradigm shift from passive AI tools to active, autonomous systems that can pursue goals, make decisions, and take actions with minimal human intervention [8]. In food safety applications, AI agents can autonomously monitor safety parameters, initiate corrective actions, and manage compliance processes [9]. Figure 3 Cloud-Based Agentic Generative AI Architecture for Food Quality Control Key applications of agentic AI in food safety include: • Autonomous quality control agents: Systems that continuously monitor production processes, make realtime adjustments, and initiate corrective actions without human intervention [10] • Compliance management agents: AI systems that autonomously track regulatory changes, update compliance requirements, and ensure ongoing adherence to safety standards [19] • Supply chain monitoring agents: Autonomous systems that track products through the supply chain, monitor environmental conditions, and detect potential safety issues [24] The policy framework for Agentic Generative AI must be designed not only to ensure robust governance and risk management of AI models [33, 34], but also to directly enhance system accessibility, affordability, and efficacy in specialized critical domains such as infectious disease management [35, 36]. The development of agentic AI systems represents a move toward fully autonomous food safety management, where AI systems not only identify issues but also implement solutions in real-time [25]. 6.3. FDA’s AI Initiatives and Regulatory Perspectives The U.S. Food and Drug Administration (FDA) has recognized the transformative potential of AI in food safety and has launched initiatives to incorporate AI technologies into its regulatory operations. The FDA’s agency-wide AI tool, "Elsa," represents a significant step toward leveraging AI for regulatory optimization and enhanced public health protection [11]. This regulatory adoption of AI technologies signals growing acceptance and validation of AI approaches in food safety. It also establishes frameworks for evaluating and approving AI-based food safety systems, creating pathways for broader industry adoption [26]. World Journal of Biology Pharmacy and Health Sciences, 2025, 23(03), 303-315 309 7. Generative AI and AI Agents: The Next Frontier The integration of Artificial Intelligence within the food industry is evolving beyond analytical and predictive models into the realms of creation and autonomous action. This evolution is primarily driven by two transformative technologies: Generative AI and AI Agents. These technologies are moving from conceptual frameworks to practical tools, offering novel solutions to long-standing challenges in food safety and quality control. Figure 4 Policy Framework for Government Implementation of Agentic Generative AI in Food Safety 7.1. Generative AI: Beyond Prediction to Creation Generative AI (GenAI) represents a significant leap from traditional AI. While conventional models are designed to recognize patterns, classify data, and make predictions, generative models can create new, original content—including text, images, strategies, and even synthetic data [5]. In the context of food safety, this creative capability unlocks several powerful applications: • Enhanced Predictive Scenario Planning: GenAI can simulate a vast array of potential contamination events, supply chain disruptions, or equipment failures. By generating these scenarios and modeling their outcomes, food safety teams can develop more robust and comprehensive mitigation and response plans, moving beyond historical data to prepare for novel threats [22]. • Automated and Dynamic Documentation: The burden of maintaining Hazard Analysis and Critical Control Points (HACCP) plans, Standard Operating Procedures (SOPs), and compliance documentation is significant. Generative AI can automate the creation and updating of these critical documents. It can analyze new regulatory updates, scientific literature, and internal process data to generate revised protocols, audit reports, and training materials tailored to the specific context of a facility [6]. • Supply Chain Optimization and Simulation: GenAI can model complex supply networks under various constraints (e.g., weather events, geopolitical issues, supplier reliability). It can generate optimal routing and inventory strategies that prioritize food safety by minimizing transit time, ensuring temperature control, and diversifying risk [7]. • Synthetic Data Generation: A major hurdle in training robust machine learning models is the lack of sufficient, high-quality data for rare but critical events (e.g., specific pathogen contamination). Generative AI can create realistic, synthetic data representing these edge cases, allowing for the training of more accurate and reliable detection and prediction models without compromising real operational or sensitive data [23]. The implementation of GenAI, however, requires careful validation. As noted by industry analysts, challenges include "a developing and unfamiliar solutions ecosystem, uncertain cost implications and the complexities of selecting the right vendor partnerships" [22]. Ensuring that generative outputs are evidence-based, scientifically valid, and aligned with regulatory requirements is paramount for its safe adoption in the high-stakes domain of food safety [27]. World Journal of Biology Pharmacy and Health Sciences, 2025, 23(03), 303-315 310 Figure 5 AI-Powered Food Quality Control Architecture with Agentic Generative AI (landscape layout, all elements preserved) 7.2. AI Agents: Autonomous Action for Continuous Assurance AI Agents represent a paradigm shift from tools that provide insights to systems that take autonomous action. An AI agent is "a type of artificial intelligence system that is capable of autonomously performing tasks and pursuing predefined goals" [8]. These agents can perceive their environment through data inputs, process information using models like large language models (LLMs), make decisions, and execute actions to achieve specific objectives, often with minimal human intervention [9]. In food and beverage manufacturing, AI agents are transforming operations by enabling stakeholders to "talk to the factory" and optimize processes in real-time [16]. Their applications are diverse and impactful: • Autonomous Quality Control Agents: These agents operate continuously, monitoring sensor data and visual feeds from production lines. They can autonomously adjust process parameters (e.g., temperature, pressure, mixing speed) to maintain product quality within specified bounds. If a critical deviation is detected that cannot be auto-corrected, the agent can initiate a shutdown, quarantine products, and immediately alert human supervisors [10], [13]. • Compliance and Regulatory Agents: These specialized agents act as autonomous compliance officers. They constantly monitor global regulatory databases, interpret new guidelines, cross-reference them with current company practices, and automatically update compliance checklists and audit protocols. They can pre-fill compliance reports and manage the documentation required for regulatory submissions, drastically reducing administrative overhead and the risk of human error [19]. • Supply Chain Monitoring Agents: Deployed across the logistics network, these agents track shipments in realtime. They monitor environmental conditions (e.g., temperature, humidity) via IoT sensors and can autonomously initiate actions if a parameter breaches safety limits. For example, an agent could reroute a shipment to a closer facility to prevent spoilage or flag it for immediate inspection upon arrival [24], [25]. The transition to agentic systems is "redefining how food is produced, processed, and delivered" [24]. By handling repetitive monitoring and decision-making tasks, they free human experts to focus on strategic oversight, complex problem-solving, and continuous improvement initiatives. This collaboration between human intelligence and autonomous AI agency is key to building more resilient, efficient, and safe food systems. 8. Policy Recommendations for Government Implementation The integration of Agentic Generative AI into food safety systems requires thoughtful policy frameworks that balance innovation with public safety. Based on the technological review presented in this paper, we propose the following policy recommendations for government agencies and regulatory bodies. 8.1. Establish AI Validation and Certification Frameworks Government agencies should develop standardized validation protocols for AI-based food safety systems. These frameworks should include: World Journal of Biology Pharmacy and Health Sciences, 2025, 23(03), 303-315 311 • Performance benchmarks: Establish minimum accuracy thresholds for AI detection systems across various food safety applications • Transparency requirements: Mandate explainable AI features that allow regulators to understand decisionmaking processes • Continuous monitoring: Require ongoing performance validation and regular updates to AI models • Third-party certification: Create accreditation programs for independent verification of AI system reliability The FDA’s development of AI tools like "Elsa" demonstrates the government’s recognition of AI’s potential in regulatory operations [11]. This initiative should be expanded into a comprehensive certification framework for industry-wide AI adoption. 8.2. Create Data Sharing and Collaboration Incentives To address the challenge of limited training data for rare food safety events, policymakers should: • Establish secure data repositories: Create government-managed platforms for anonymized food safety data sharing • Develop data standards: Implement uniform data formats and quality requirements for AI training data • Provide tax incentives: Offer benefits to companies that contribute high-quality safety data to public repositories • Foster public-private partnerships: Encourage collaboration between regulatory agencies, academic institutions, and industry partners These measures would help overcome the data scarcity issues identified in Section 9 while maintaining privacy and competitive protections. 8.3. Modernize Regulatory Frameworks for Autonomous Systems Current regulatory frameworks were designed for human-centric operations and require updating to accommodate autonomous AI systems: • Update Good Manufacturing Practices (GMPs): Revise GMP regulations to include requirements for AI system validation and maintenance • Develop AI-specific HACCP guidelines: Create guidance for integrating AI into Hazard Analysis and Critical Control Points systems • Establish liability frameworks: Clarify responsibility and accountability for decisions made by autonomous AI agents • Create adaptive regulations: Implement regulatory frameworks that can evolve with technological advancements These updates should build on existing initiatives like AI-enhanced HACCP management [18] while ensuring regulatory frameworks remain technology-neutral and outcome-focused. 8.4. Support Research and Development Initiatives Government investment in AI food safety research is crucial for advancing the field: • Fund academic research: Support university research on AI applications in food safety and quality control • Create innovation grants: Provide funding for small and medium enterprises developing AI food safety solutions • Establish testbed facilities: Fund pilot programs and testing environments for new AI technologies • Support workforce development: Invest in education and training programs for AI specialists in food safety These initiatives would address the expertise gap identified in Section 9 and accelerate the development of validated AI solutions. 8.5. Implement Graduated Compliance Timelines To ensure equitable adoption across the food industry, policymakers should implement phased compliance requirements: