Artificial Intelligence Applications in Risk Management Within Integrated Management Systems: A Review

Abstract

Artificial intelligence is increasingly used in all fields, especially in the area of risk management within Integrated Management Systems (IMS). The paper aims to highlight the role of artificial intelligence (AI) in risk management, therefore providing opportunities for industrial organizations, offering significant advantages for improving the efficiency and accuracy of risk assessment and mitigation processes. By using advanced AI technologies, organizations can anticipate and manage risks more effectively, therefore optimizing operational performance and resilience. We reviewed and explored the main applications of AI implementation, risk management, the barriers encountered, and the advantages and disadvantages of using AI. A holistic analysis of IMS risk management, identification and assessment, operational efficiency of routine tasks, real-time data analysis, and immediate decision-making using AI was performed. The methods and technologies used are analyzed, along with the associated challenges, providing a comprehensive perspective on the impact of AI in industrial organizations. We conclude that the use of AI addresses challenges related to data quality, model interpretation, ethical issues, and high costs of implementation and management, which require qualified personnel. Also, we conclude that the use of AI in risk management for IMS presents significant opportunities for industrial organizations, including enhanced process monitoring, rapid information analysis, and swift response to emerging risks. This enables the optimization of risk management strategies, ultimately leading to increased operational safety and efficiency.

1. Introduction

Artificial intelligence (AI) has evolved significantly in recent years, having a profound impact on society as a whole across all fields of activity. The roots of AI can be traced back to the late 1940s, when pioneers in the field recognized that, by implementing the right algorithms, computers could perform tasks that mimic human intelligence. An early example of this is the checkers program created by Christopher Strachey in 1952, which was later optimized by Arthur Samuel, integrating machine learning techniques [1].

Since 2000, AI has undergone rapid growth, significantly influencing the development of human society through revolutionary advancements in theories and techniques [2]. This growth has been supported by advances in hardware and software, powerful computational platforms, and access to massive amounts of data [3]. AI has become a complex interdisciplinary field with applications in various sectors, including health, industry, transportation, agriculture, and entertainment, etc. [4,5].

While AI offers significant opportunities for innovation and efficiency, it also presents challenges related to complexity and the reliance on larger volumes of data [6]. Market success in the AI domain is contingent upon the co-evolution and integration of technology with novel business models to establish a dominant design [7]. Additionally, analyzing the emergence of technology and technological opportunities in AI can provide valuable insights for future development [8].

The integration of artificial intelligence into Integrated Management Systems (IMS) represents an essential paradigm shift, influencing business models, processes, and the competitiveness of companies in the industry [9].

A management system (MS) is an organized and structured approach to managing an organization or a department within the organization, which requires setting objectives, developing processes, instructions, procedures, monitoring and evaluating performance, and continuous improvement, for one or more MS, such as the Quality Management System (QMS), the Environmental Management System (EMS), the Occupational Health and Safety Management System (OHSMS) [10].

In Germany, the use of AI is already considered a core technology that can provide a significant competitive advantage through its disruptive use in management [11].

The strategic use of AI in management can bring significant benefits, such as decision support, customer and employee engagement, process automation, and the development of new products and services. However, there are challenges related to the practical use and strategic integration of AI to create business value [12], help manage the business, and take into account the new standard form IMS.

The emergence of the ISO/IEC 42001:2023 standard [13] provides a framework for the responsible management of artificial intelligence in organizations. It is applicable regardless of the organization’s size or sector and is designed to integrate with existing standards supporting a holistic approach to AI governance.

Integrating AI into management systems requires a well-defined methodological approach. This includes shaping organizational culture, setting goals, identifying performance indicators, and integrating AI products into the management system. To ensure the success of the implementation, it is essential to monitor results and audit the management system [14]. AI has become a disruptive force in task management systems, improving process planning, prioritizing needs, and completing tasks in a very short time. Modern platforms use AI to automate repetitive tasks and provide intelligent recommendations, facilitating intuitive interaction for proactive risk management [15]. The use of AI in strategic and operational management can improve decision-making efficiency and innovative activity. A strategic approach is essential not only to optimize management processes for competitive advantage but also to mitigate the unique risks inherent in AI implementation [16]. Using AI in the management system provides the opportunity to better manage the organization, thereby improving all operational processes, both in production and in enabling top management to make immediate decisions regarding decision-making processes. The realization of these successes is predicated on two key prerequisites: the deployment of an optimal strategy and the proactive addressing of inherent challenges, notably organizational change resistance and the imperative to ensure ethical integrity and data confidentiality.

2. Research Methodology

2.1. Methodology

The research methodology in a literature review article refers to the systematic approach used to address research questions and synthesize existing knowledge. It comprises selecting appropriate methods and techniques, ensuring that the research process is rigorous and relevant.

The chosen methodology should be aligned with the specific research problem, as distinct issues may necessitate differing methodological approaches [17].

In undertaking this research, the aim was to formulate and investigate fundamental questions that address the issue of using artificial intelligence in the process of managing risks associated with IMS, as follows:

• What are the strategies implemented for using AI in IMS risk management?
• What challenges arise in integrating AI into IMS?
• What are the opportunities generated by implementing IMS using AI?
• What are the advantages of integrating artificial intelligence into IMS in industrial organizations?
• What are the risks associated with using AI?
• What are the barriers to using AI in IMS risk management?
• Are Internet of Things (IoT) devices and technologies suitable for data collection and IMS risk management using AI?
• What are the main emerging risks encountered when using AI?
• What are the critical factors in IMS risk management using AI?
• What algorithms and technologies are used for risk management of IMS using AI?
• How is IMS risk management using AI achieved in the management of documents developed by industrial organizations?

To answer the questions listed above, field literature was analyzed, focusing on recent research on the application of AI in IMS.

In Figure 1, the bibliographic sources used to carry out the research were grouped according to interest as follows: Risk and safety management, other/general AI applications, production and Industry 4.0, enterprise resource planning (ERP) and integrated systems, construction and civil engineering, applications for finance and business, education and research methodology and ethical regulation of AI.

A distribution of bibliographic sources was made according to the year of publication of the works (Figure 2).

To obtain a comprehensive perspective on the current state of knowledge in the field, a systematic review of the literature was conducted using major databases: MDPI, Elsevier, SpringerLink, Emerald Insight, Taylor & Francis, and Google Scholar. The review was conducted between January 2025 and September 2025.

The first stage of the study involved searching six databases for relevant articles using the following keywords: ‘risks’, ‘integrated management systems’, ‘opportunities’, ‘industrial organizations’, and ‘digital technologies’.

After identifying the works relevant to the proposed topic, a selection of duplicate titles was made using the Excel program, through the Home → Conditional Formatting → Highlight Cells Rules → Text function, based on the information collected from the six databases. The analysis phase, subsequent to the identification of relevant documents (Figure 2), involved systematically extracting data to address the study’s core objective: answering 11 foundational questions on the role of artificial intelligence in the risk management processes of integrated management systems (IMS).

2.2. Bibliometric Analyses

After all the necessary documents were collected, the VOSviewer version 1.6.19 software was used to perform the bibliometric analysis. To perform the bibliometric analysis, papers were selected from platforms such as MDPI, Elsevier, Taylor & Francis, Emerald Insight, Springer, and Google Scholar, using filters for specific periods and topics, as shown in Figure 3 (keyword analysis) and Figure 4 (links between paper titles, keywords, and abstracts).

The bibliometric analysis presented in Figure 3 illustrates that the literature on artificial intelligence is structured in three interdependent directions: responsible development and risk management, organizational impact and theoretical models, and practical applications in industry and technology. This complementarity reveals the field’s evolving dialectic tension, characterized by the interplay of caution and innovation, as well as conceptual construction and practical implementation.

The co-occurrence analysis of keywords was performed using the VOSviewer v1.6.19 software. A minimum frequency threshold of 8 occurrences (≥8) was established for the inclusion of terms in the analysis. The similarity between terms was calculated based on the association strength, a method that normalizes co-occurrences to highlight the real intensity of conceptual relationships. The grouping of terms was performed using the modularity-based clustering algorithm, which maximizes the internal cohesion of each thematic cluster. The size of the nodes in the network reflects the frequency of occurrence of each term, and the thickness of the connecting lines indicates the strength of the co-occurrence relationships. The different colors correspond to the thematic clusters automatically identified by the software.

Figure 3 shows the co-occurrence network of keywords extracted from the bibliometric map. It highlights three major thematic clusters corresponding to the main research directions. The clustering is performed based on similarity in all figures created with the VOSviewer software.

The green cluster groups terms such as risk, risk management, development, and use, representing the field of responsible development and risk management associated with AI technologies.

The red cluster includes terms such as impact, model, organization, and innovation, suggesting an orientation towards theoretical models, organizational impact, and innovation processes.

The blue cluster integrates terms such as application, IoT, thing, manufacturing, and machine learning, reflecting the area of practical applications in industry and technology.

The thickness of the lines connecting the nodes indicates the strength of the co-occurrence relationships. The density of the network highlights the high level of interdisciplinarity between the fields and demonstrates the interconnection between the concepts of risk management, innovation, and application.

Therefore, the integration of AI cannot be approached unilaterally but requires a systemic vision capable of combining risk analysis, scientific validation, and practical applicability (Figure 4). For this map, 3379 items grouped into 19 clusters were analyzed.

Representation of the co-occurrence network, allowing the observation of the thematic evolution of the AI field. The color of each node indicates the average year of publication of the publications in which the respective term is predominant.

The analysis shows a gradual shift from conceptual and theoretical themes, represented by shades of orange and green and linked to the years before 2018, to recent technological applications, shown by shades of red, yellow, and blue and associated with 2020–2024. The early period is dominated by terms like risk management and development, while concepts such as the Internet of Things (IoT), machine learning, and applications have emerged more recently. This illustrates the evolution of research from conceptual models to practical implementations in industry and technology. This distribution confirms the current trends in the field, focusing on the intelligent integration of AI technologies into risk management processes and the development of a transdisciplinary approach.

The varying thickness of the connecting lines in Figure 4 reflects the frequency of co-occurrence of terms between the three major clusters identified—“responsible development and risk management” (green), “organizational impact and theoretical models” (red/yellow/brown), and “practical applications in industry and technology” (blue/pink)—showing that these themes are closely interconnected in the AI literature, highlighting an interdisciplinary framework in which the concepts of responsibility, theory, and practical applicability are frequently analyzed together. The cluster settings for Figure 4 are as follows: random stars 5, Iterations 5, random seed 2, resolution 3, and minimum cluster size 5.

In Figure 5, the links between terms represent the frequency of their co-occurrence in the scientific literature, and the density and complexity of the network highlight a high degree of thematic interconnection in research on artificial intelligence in the last 4 years.

The strongest and most numerous connections start around the central term artificial intelligence, indicating that it is approached in close relation to topics such as process, management, risk, data, and application. These links reflect not only the simultaneous use of these concepts in the same studies but also their integration into a multidisciplinary approach, in which technological, organizational, and strategic aspects are analyzed together. At the same time, the presence of connections between terms in different areas of the graph emphasizes that the literature does not treat subtopics in isolation, but pursues a holistic approach to the impact and implementation of artificial intelligence.

The prevalence of keyword connections established within the last four years, as shown in Figure 5, reflects a significant recent intensification of research activity. This clustering suggests the emergence of new, dynamic research fronts that present timely opportunities for exploration. This temporal distribution of co-occurrences confirms the current and emerging nature of the topic regarding the application of artificial intelligence in risk management within integrated management systems.

In conclusion, the literature on the use of artificial intelligence in IMS risk management is structured around four main pillars: risk management, organizational implementation, efficiency enhancement, and technological applicability in industry and IoT. The results show that AI is not approached unilaterally, but as an interdisciplinary field in which safety, performance, and innovation are interdependent. Contemporary research trajectories reflect the tension between harnessing the transformative potential of artificial intelligence and addressing the imperative of responsibly managing its associated risks.

3. Results and Discussion

Integrating AI into risk management systems has become a core practice across industries, including finance and healthcare. AI offers significant advantages in identifying, assessing, and mitigating risks, improving decision-making and operational efficiency [18,19,20]. AI enables rapid and accurate analysis of large volumes of data, facilitating early identification of risks and improving predictive capabilities [21,22]. In the financial sector, AI has revolutionized credit risk assessment and market analysis, delivering a 20% increase in predictive accuracy and a 60% reduction in false alarms in fraud detection [22,23]. In hospitals, AI contributes to patient safety by detecting critical conditions early on and improving clinical risk management [18].

The use of artificial intelligence in IMS risk management marks a major shift from traditional methods and is a trending topic with the potential to transform risk management through automation and optimization. AI offers enhanced predictive capabilities and operational efficiency, enabling the identification, evaluation, and mitigation of organizational risks in real-time.

In the era of digitalization, the use of AI in IMS has become a necessity to enhance operational efficiency and decision-making processes within any organization. The significant evolution of Industry 4.0 has necessitated a fundamental shift in how organizations manage their processes and risks, compelling the adoption of multifaceted strategies [24,25,26,27]. In

Table 1. Strategies for using AI in IMS risk management.

No.	Name of the Strategy for Implementation	Description of the Strategy	Authors
1.	Using a dashboard to track AI progress	There must be a document that clearly shows the evolution of integrating AI into risk management. This should cover all stages, from raising awareness of the need for AI, to carrying out the necessary analyses and preparing for implementation, and finally ensuring the system is functioning properly.	[28,29]
2.	Automatic extraction of risk factors from large volumes of textual data	Sing NLP (Natural Language Processing) algorithms to analyze documentation, audit reports, and legislation to update the risk matrix	[30]
3.	Using AI technologies for monitoring and prediction	In industries such as oil and gas, AI can be integrated into health, safety, and environment (HSE) management systems to improve real-time monitoring and predictive risk management. This enables proactive risk identification and mitigation.	[31]
4.	Automating and optimizing project management processes	AI can optimize resource allocation and mitigate risks in project management through big data analysis and scenario simulation.	[32,33]
5.	Integration into ERP projects	AI can be used to enhance the risk assessment methodology employed in ERP projects, enabling managers to respond more swiftly to emerging threats.	[34]
6.	Integration with IoT and Big Data	AI can correlate data from sensors and monitoring equipment, as well as human reports, to provide smart monitoring of identified environmental aspects. It can also monitor hazards and factors affecting employee health and safety.	[35]
7.	Risk assessment in agile projects	In IT projects managed using the Scrum methodology, AI can improve the accuracy and depth of risk assessments, enabling proactive risk management.	[36]
8.	Assessing the risks of implementing AI	Using an appropriate taxonomy in identifying and classifying risks that arise in AI systems.	[16]

, this study identifies several strategies for leveraging AI in industrial organizations to enhance risk management within an IMS.

Recent studies highlight that the use of AI within IMS can increase operational efficiency by up to 40% [37].

The use of AI in management can be difficult due to several challenges. These include the complexity of management phenomena, constraints related to data sets, questions of accountability related to fairness, and other ethical and legal constraints, as well as adverse employee reactions to AI-based management decisions [38].

The challenges that arise when implementing AI in IMS are diverse and may include:

• Systems management: Conflicts in integrating AI with existing systems, including compatibility and complexity of existing IMS architecture [25,39,40].
• Quality, complexity, and existing information: Implementing processes in IMS is complex, and insufficient or sparse data sets can limit the effectiveness of AI [38,41].
• Ethical issues and legal requirements: There are concerns about fairness, privacy, and transparency, which can complicate the implementation of AI [38,42,43].
• Technological integration: Integrating AI with existing IT infrastructures can be difficult, requiring compatibility and effective collaboration [42,44] and additional costs [41].
• Skills shortage: The shortage of qualified personnel in AI technologies is a major challenge, requiring investment in training and development [25,44,45,46].
• Resistance to change: Cultural and organizational resistance to change can hinder AI adoption [38,46].

The integration of Artificial Intelligence into IMS implementation presents significant strategic opportunities for organizations, such as enhanced operational efficiency, reduced costs, and improved customer satisfaction. Using AI to integrate management systems can be more beneficial than implementing them individually, as it allows for a more comprehensive approach [47,48]. The opportunities driving the use of AI in IMS management are presented in

Table 2. Opportunities generated by implementing IMS using AI.

No.	Opportunity	Details	Authors
1.	Increasing the efficiency and quality of decisions	AI allows for the optimization of data and information flows used in decision-making and can improve decision-making processes by reducing time	[49,50]
2.	Decision support	AI provides decision support by identifying patterns and anomalies in data	[44,49]
3.	Improving operational efficiency	AI can automate routine tasks, optimize workflows, leading to increased operational efficiency, driving increased productivity and economic efficiency	[32,44,49]
4.	Cost reduction	Implementing AI can reduce operational costs by optimizing resources and automating processes	[44,49]
5.	Transforming resource management	AI transforms resource management practices through efficiency and innovation	[51,52]
6.	Supply chain optimization	AI improves demand forecasting, inventory management, and logistics optimization	[53,54,55]
7.	Strategic innovation	AI drives innovation by integrating advanced technologies such as machine learning and natural language processing	[55,56]
8.	Innovation and continuous improvement	AI identifies new opportunities for product development and innovative processes	[57,58,59,60]
9.	Improving customer satisfaction	AI can improve customer satisfaction through faster and more accurate decision-making processes	[49]
10.	Sustainability and environmental management	AI contributes to sustainable practices by optimizing resource use, real-time monitoring for rapid identification of deviations, proposing solutions to reduce environmental impact, and reducing emissions and waste	[24,51,53]
11.	Decision support and IMS risk management	AI allows for the identification of trends and the anticipation of operational, environmental, or OHSAS risks	[26,28,50,61,62]

.

Integrating AI into IMS in industrial organizations fundamentally transforms the way organizational processes are managed. The benefits are multifaceted, from operational efficiency and resource optimization to improved safety, quality control, and predictive maintenance. In addition, this integration supports data-driven decision-making, adaptability to change, and continuous innovation, thus ensuring a long-term competitive advantage for industrial organizations, ultimately satisfying stakeholders.

The implementation of these modern IT technologies reduces costs and enhances productivity while simultaneously fostering a safer and more adaptable work environment, thereby constituting an essential phase in modern industrial evolution. Some of the advantages of integrating AI into IMS are described in

Table 3. Advantages of integrating artificial intelligence into IMS in industrial organizations.

No.	Advantages	Description of the Obtained Advantages	Authors
1.	Predictive maintenance	Machine learning algorithms minimize downtime and optimize resource allocation.	[63,64,65,66,67]
2.	Operational optimization	Real-time data analysis improves decision-making and resource utilization.	[65,68,69,70,71,72]
3.	Quality control	Advanced image recognition and machine learning ensure higher standards.	[63,65,68,71,73,74,75]
4.	Human–machine Collaboration	Collaborative robots and AI are improving production capabilities.	[63,76,77,78]
5.	Process automation	Reducing human intervention increases product efficiency and innovation.	[78,79,80,81]
6.	Accelerated financial analysis	AI improves budgeting accuracy and reduces planning time.	[65,82]
7.	Productivity increase	Optimizing production processes and reducing equipment downtime.	[68,75,77,80,82]
8.	Improving customer relations	Optimization of customer relationships and business processes.	[65,82]
9.	Automation of logistics	AI improves production logistics and supplier interaction management.	[80,83]
10.	Supply chain management	AI predicts demand, optimizes inventory, and streamlines routes.	[26,63,65,72,84]
11.	Data-driven decisions	AI facilitates informed decision-making and agile response to market conditions.	[67,68,72]
12.	Reducing inventory management costs	AI optimizes inventory management and reduces costs.	[67,82]
13.	Improving transparency	AI increases transparency in supply chain processes.	[67,72]
14.	Increasing flexibility	AI enables smarter and more flexible systems in production.	[67,71]
15.	Improving cross-functional collaboration	ERP integration facilitates collaboration and real-time data exchange.	[68,72]
16.	Intelligent process planning	AI assists intelligent process planning through deep learning.	[59,68,75,78,83]
17.	Increasing customer satisfaction	Higher quality standards improve customer satisfaction.	[67,73,74,81]
18.	Error reduction	AI improves accuracy and reduces errors at rates beyond human capacity.	[78,80]
19.	Increasing sustainability	AI contributes to the creation of sustainable industrial value.	[66,68,75,80,85]
20.	Improving data security	AI requires robust security elements to protect data.	[78,86,87]

.

Using AI in IMS risk management brings both benefits and significant risks. A description of the main risks associated with using AI in this context is detailed in

Table 4. Risks associated with using AI.

No.	Description of the Risk	Detailing the Risk	Authors
1.	Impact on organizational culture and employees	The use of AI can transform the organization by changing structure and authority; it can introduce additional complexity, which can increase resistance to change.	[88,89,90]
2.	Errors in data processing and decision-making	AI can generate wrong decisions if the input information is incorrect, incomplete, or inaccurate.	[40,61]
3.	Lack of transparency in data processing	AI can operate as a “black box”, making it difficult to understand how decisions are made; AI algorithms can be designed or trained on biased data.	[91,92]
4.	Cybersecurity impact	AI systems can be vulnerable to cyberattacks, which can compromise data and system operation.	[26,92,93,94]
5.	Inadequate implementation	AI can be a source of risks depending on the existing variant, the lifecycle of the IMS, and the interaction with it.	[61,93,95]
6.	The quality of the data used	Insufficient data, of poor quality, unreliable, and with an uncertain prediction.	[25,45]
7.	Algorithmic bias	AI algorithms can perpetuate or amplify existing biases or errors in training data, affecting the accuracy of decisions.	[93,95,96,97]
8.	Privacy issues	The use of AI can lead to data privacy breaches, especially if data is used without consent.	[97,98]
9.	Reduced human factor supervision	Excessive automation can reduce human oversight, increasing the risk of unnoticed errors.	[97]
10.	Difficulties in managing large volumes of data	AI requires processing large amounts of data, which can be difficult to manage.	[20]
11.	High implementation costs	Implementing and maintaining AI systems can be expensive.	[20,41]
12.	Ethical challenges	Automated decisions carry ethical risks that can have negative impacts on individuals and society.	[43,97]
13.	Uncertainties in data manipulation	There are risks related to the incorrect manipulation and interpretation of data.	[92,99]
14.	Ensuring compliance with legal and regulatory requirements	Limitations in the design and sources of AI training can lead to non-compliance with legal and reputational impacts.	[100,101]
15.	Regulatory challenges	Lack of clear regulations can lead to inappropriate use of AI.	[23,92]
16.	Excessive dependence on technology	Decisions in the organization may neglect human intervention that requires the use of critical skills to analyze processes.	[54,102]

.

In managing the risks of using AI, several barriers are encountered that must be carefully analyzed to establish certain resilient strategies to counteract these barriers (

Table 5. Barriers encountered when using AI in IMS risk management.

No.	Type	Category	Authors
1.	Technological barriers	Quality and availability of data provided a. Lack of a standardized collection and storage system; b. Lack of data standardization, which complicates their processing and analysis by AI algorithms; c. Difficulties in integrating relevant historical data; d. Recording errors; e. Confidentiality.	[20,26,93,103,104,105]
		Use of hardware, software, and data infrastructure a. Lack of MLOps platforms for “controlled” and auditable operation. b. The hardware is outdated and cannot support advanced computational algorithms; c. Lack of interoperability between different systems; d. Need for investments in IT infrastructure; e. AI integration requires an infrastructure with advanced resources, quality input data, and functional and robust processes.
		Lack of transparency of computational algorithms a. Users’ distrust of AI recommendations; b. Ability to justify certain decisions to auditors or authorities; c. AI models are often opaque, which makes it difficult to audit, validate, and accept decisions; d. Lack of human validation of some decisions.
2.	Organizational barriers	Lack of AI alignment with risk management objectives: a. Unclear objectives; b. Difficult mapping of processes and performance indicators; c. Management of AI-assisted decisions; d. Failure to assign responsibilities to the process owner; e. Lack of sufficient evidence for SMI audit and associated risks.	[20,26,104,106]
		Lack of competencies in using AI and of specialists in the field a. Data science and machine learning; b. Cybersecurity; c. Risk management and auditing.
		Integrating AI with existing processes: a. Organizations use traditional management methods using ISO 9001 [107], ISO 14001 [108], ISO 45001 [109], etc. b. Risk management related to bias, transparency and accountability; c. Over-reliance on AI for environmental and OHSAS hazard identification.
3.	Legal, ethical and, security barriers	Regulatory compliance: a. There are no standards for auditing and validating AI systems b. Regulations on personal data protection; c. Regulations in certain areas such as transport, energy, pharmaceuticals, etc.; d. Ensuring legality regarding documented information control and confidentiality requirements.	[18,93,104,106]
		Ethical and liability issues: a. Potential for discrimination due to data bias; b. Lack of control over decision-making using AI; c. Monitoring employee behavior.
		Cybersecurity: a. Data poisoning; b. Attacks on hardware, software, and data infrastructure running AI.
4.	Economic and resource barriers	Developing and implementing AI requires a. Purchasing hardware, software, data, consulting, and staff training; b. Purchasing licenses; c. Costs for integrating existing systems with AI; d. Infrastructure changes; e. Costs of specialized personnel; f. Uncertain return on investment.	[22,23,104,110]
		Maintenance and update costs a. AI must be constantly updated to be able to detect potential risks.
5.	Human barriers	Resistance to change a. Conservative organizational culture, b. Employee reluctance to use AI; c. Dilemma of ensuring job satisfaction; d. Concerns about new responsibilities and work tasks; e. Fear of replacement by AI; f. Communication and collaboration issues.	[103,104,111,112,113]
		Uncertainties a. Lack of trust in the recommendations or decisions of an AI system; b. Overconfidence leading to loss of vigilance and critical analysis skills; c. Responsibility and accountability regarding the use of AI.

). The implementation of AI for risk management in IMS faces multifaceted barriers of a complex nature: technological, organizational, human, legal, ethical, data security, integration, and cost. Successful integration of AI into an IMS is contingent upon establishing a triad of foundational elements: appropriate policies, a strategic approach dedicated to transparent processes, and standardization for interoperability with existing systems.

An AI-driven solution for IMS risk management relies on a network of essential IoT devices that provide the real-time data collection and analytical capabilities required for effective risk management within an industrial enterprise. Using AI and other emerging technologies can significantly improve the efficiency and security of industrial operations, helping organizations become more resilient.

Table 6. IoT devices and technologies for data collection and IMS risk management using AI.

No.	IoT Devices and Technologies for Data Collection and IMS Risk Management	Detailed Description of IoT Devices and Technologies for Data Collection and IMS Risk Management	Authors
1.	IoT and cloud-based monitoring platforms	Utilizing AI techniques such as auto-encoders and fuzzy inference systems, these platforms facilitate comprehensive production system health assessment and anomaly detection.	[114]
2.	IoT devices with semantic management	Industrial IoT devices use on-device applications to analyze data in real time, facilitating decision-making at the edge through techniques like TinyML and complex event processing.	[115]
3.	Operational and Prognostic Management (PHM) Systems	AI-based PHM technologies are used for condition monitoring, failure prediction, and equipment operability management in various industries.	[116]
4.	Edge computing with AI	Integrating AI into edge computing improves the computing speed and energy efficiency of industrial IoT devices, making it essential for monitoring industrial platforms.	[117]
5.	Intelligent IoT architectures with blockchain	IoT architectures that integrate blockchain and AI enable efficient big data analysis and ensure data security and privacy.	[118]
6.	Anomaly detection algorithms	Deep learning algorithms to detect anomalies in industrial IoT data are used, improving the accuracy of predictions and reducing the risks of cyberattacks.	[119]
7.	AI-based HSE management systems	These systems use AI to monitor safety in real-time and manage risks proactively in industries such as oil and gas.	[31]
8.	Digital Twin Systems	Digital Twin and Generative AI technologies are used to simulate and manage assets in cyber-physical production environments.	[120]
9.	Private AI Frameworks for Industrial IoT	Private AI frameworks combine continuous learning and encryption to ensure data and model security in industrial IoT environments.	[121]
10.	Machine Learning-based automation platforms	Automation platforms use AI and big data technology to inform engineering project decisions, thereby improving risk management and work accuracy.	[122]

presents IoT devices and technologies for data collection and IMS risk management.

The use of IoT devices and technologies for data collection and IMS risk management, in turn, creates new emerging risks that must be thoroughly analyzed to ensure sustainable development and efficiency of organizations. The main emerging risks identified are:

• Opacity and inaccuracy: AI can introduce risks of opacity and inaccuracy into risk assessments, which can lead to operational inefficiencies and incorrect risk assessments [92].
• Cybersecurity and data manipulation: There are significant cybersecurity risks and uncertainties regarding data manipulation, which can compromise the integrity of management systems [92,93].
• Loss of human control: The rapid development of autonomous AI systems may lead to the loss of human control, with potential negative social and economic consequences [123,124].
• Biases in algorithms: AI algorithms can contain biases that affect decisions and risk assessments, requiring a responsible framework to manage these issues [92,125].
• Monitoring challenges: Effective monitoring of AI systems is difficult, which can lead to unforeseen risks and system failures [92,93,126].

The expanding influence of AI, marked by its advanced analytical and predictive capabilities, presents a dual reality: while its capacity for continuous improvement offers significant promise, it simultaneously introduces a suite of emergent risks that demand careful management. This leads to ongoing emerging risks. The evolution of AI must be continually monitored, as AI can introduce instability and uncertainty to organizations. The lack of clear regulations and transparency in how these systems operate can create uncertainties in data handling and monitoring, increasing cyber threats and algorithmic biases. Furthermore, the use of AI in risk management can result in information overload, thereby complicating the decision-making process [92].

Fundamental differences between AI systems and classic software create new sources of risk that must be identified and managed early to prevent subsequent systemic failures [93]. Rapid progress in AI, particularly within autonomous systems, carries a dual peril: it could exacerbate existing social risks while simultaneously diminishing direct human control [123].

Within the domain of industrial organization and operations management, AI is being leveraged to enhance supply chain resilience and optimize enterprise risk mitigation frameworks. However, this use requires a careful approach to avoid costly decisions and manage demand uncertainties [127]. Also, in the construction field, AI contributes to digital transformation, but requires a thorough understanding of the associated technological and ethical risks [126].

While AI presents significant opportunities for organizational improvement through enhanced risk management, its implementation must be both responsible and adaptive. This approach is necessary to address emerging risks and ensure a sustainable, ethical, and secure integration, thereby mitigating the unique challenges inherent in AI-enabled IMS. Some critical Factors in IMS risk management using AI are presented, as follows:

• Integrated risk approach: It is essential to consider the synergistic effects of risks when integrating multiple management systems, not just from a single perspective [128,129,130].
• The quality of managerial decisions: Management decisions must reduce risks to an acceptable level and ensure the achievement of the organization’s objectives [127,131,132].
• Critical success factors (CSF): Identifying and implementing CSFs are essential for the success of risk management systems, including factors that influence the implementation, design, and administration of systems [133,134].
• Resources and interactions: Proper resource allocation and efficient interaction between departments are crucial for risk prevention [118].
• Integrating control and risk management: IMS must connect managerial control with risk management to correctly assess the implications of decisions on the value and rating of organizations [130,132].
• Technical and social factors: Risk assessment should include both technical (e.g., equipment failures) and social (e.g., regulations, organizational factors) factors [135].
• Risk assessment and planning: Risk assessment must be carried out in stages and the results integrated into the planning of activities [136,137,138].

For data collection with IoT devices and technologies, certain algorithms use AI to identify and manage risks in the IMS of industrial organizations. These systems are essential for optimizing resource allocation, mitigating risks, and improving decision-making in complex environments.

Significant algorithms and technologies are used for risk management of IMS using AI, as:

• Decision support systems: They use AI to integrate causal and predictive networks, multi-criteria analysis, and knowledge engineering techniques for industrial risk management [32].
• Learning automation platforms: Platforms like the engineering machine-learning automation platform (EMAP) use advanced AI/machine learning (ML) algorithms to support decisions at every stage of engineering projects, from bidding to maintenance [122].
• AI-enhanced project management systems: They optimize resource allocation and risk mitigation through big data analysis and decision-making automation [33].
• The use of algorithms and technologies for risk management of IMS using AI offers considerable advantages for optimizing resources in the field of industrial organizations, such as:
• Real-time monitoring and prediction: AI enables continuous monitoring of hazardous conditions and predictive analysis of historical data to anticipate risks [31].
• Risk management in supply chains: AI improves supply chain agility by modeling scenarios and optimizing decisions [127].
• Process automation and optimization: Integrating AI with other soft computing methods, such as artificial neural networks, improves accuracy and efficiency in construction and risk management [139].

The implementation of AI algorithms is primarily limited by several key factors. Foremost among these are data quality and model interpretability; the former is fundamental to algorithmic performance, and the latter is crucial for generating transparent and trustworthy decisions. A third significant challenge is the technical and organizational difficulty of integrating these AI systems within existing technological infrastructures. Integrating AI with existing management tools can be complex and requires significant investments in infrastructure and training [33].

AI-driven algorithms for IMS risk management provide sophisticated capabilities for resource optimization and risk mitigation within industrial organizations. Although there are challenges related to data quality and system integration, the benefits in terms of efficiency and accuracy make AI a valuable tool in managing complex projects [140].

Based on the above text, in Figure 6, the steps for the strategic implementation of AI in risk management for an IMS (QMS-EMS-OHSAS) are presented.

The industrial organizations develop documents using AI in IMS. In managing the risks associated with documents developed in an industrial organization, new sources of risk generated by the use of AI are significantly influenced by current government regulations and standards. These risks arise from the integration of AI into various processes and systems, and addressing them effectively requires adapting mitigation measures and ensuring compliance with new emerging requirements. Specific AI Risk Sources:

• Operational and systemic risks: AI can introduce operational and system risks, such as errors in automated processes or failures of safety systems, which require careful assessment and proactive management [93,141].
• Safety risks and discrimination: The use of AI in industrial and human resources environments can lead to safety and discrimination risks, which must be managed through technical standards and legal regulations [142].
• Social and legal risks: AI can affect fundamental rights and create social and legal risks, requiring a regulatory approach that includes ongoing stakeholder participation and a contextual risk assessment [143].

To support the implementation of AI, some regulations and standards are going to be developed and implemented in the domain:

• EU AI Act: Proposes a risk-based approach to AI regulation, setting requirements for high-risk AI systems and promoting compliance with ISO standards [144].
• International and European Standards: play a crucial role in defining technical requirements for developing and testing AI systems, helping to mitigate risks [142].
• AI risk profiles: Proposing standards for disclosing AI risks before deployment helps inform procurement and deployment decisions and guide regulatory frameworks [145,146].

These regulations are at the beginning of their implementation, and it seems that this aspect will be developed. Managing the AI risks in industrial and corporate organizations requires continuous adaptation to new regulations and standards. An integrated approach that includes assessing operational, safety, social, and legal risks, alongside compliance with emerging regulations such as the EU AI Act and international standards, is essential. This will ensure the ethical and safe use of AI technologies [147,148,149].

The use of AI in IMS risk management represents a major step in the digital transformation of modern organizations. While AI offers revolutionary potential to optimize processes, boost efficiency, and refine strategic decisions in the interconnected contexts of Industry 4.0 and the global economy, recent studies underscore that its significant benefits are accompanied by profound challenges, necessitating a rigorous and responsible implementation framework. The integration of AI into IMS risk management yields direct and substantive benefits, including enhanced resource management, superior quality control, predictive maintenance capabilities, and more robust financial analysis.

An important aspect is the ability of AI to analyze large volumes of data in a very short time, which facilitates early identification of risks and leads to informed decision-making. AI facilitates data-driven decision-making, transforms supply chains into dynamic and intelligent networks, and improves customer satisfaction. Process automation through AI leads to cost savings, reduced processing time, and increased quality of products or services offered.

While the use of AI in IMS risk management offers significant benefits, it also poses several important challenges. These include technical difficulties in integrating with existing IT infrastructure, skills gaps among employees, resistance to change, and the complexity of the data used. AI also raises serious ethical and legal questions: the transparency of algorithmic decisions, the protection of data confidentiality, and the risk of perpetuating bias in automated decision-making.

The studies reviewed identify several risks associated with AI, including a lack of transparency, cyber vulnerabilities, algorithmic bias, and regulatory challenges. These risks emphasize the urgent need for clear regulatory frameworks, thorough algorithm testing, and human oversight to manage dangers, prevent negative impacts, and maintain a balance between automation and human responsibility.

Conversely, the opportunities that AI offers for managing the risks of IMS implementations are numerous and consistent. Among the most relevant are: supporting strategic decisions by identifying hidden patterns, supporting innovation by integrating emerging technologies, reducing operational costs through automation, optimizing the supply chain, and contributing to sustainability by reducing waste and resource consumption. In this sense, AI becomes a central pillar of digital transformation and sustainable organizational development. In a constantly changing global environment, AI can become the main driver of innovation and performance, provided that it is implemented in a responsible, sustainable, and tailored way for each organization.

Having clear strategies for implementing AI is essential to capitalize on these opportunities. Examples in the literature, such as the thorough documentation of steps for applying AI to risk monitoring and analysis in critical industries, or the use of AI in ERP projects, demonstrate that the successful implementation of AI depends on detailed planning and a strong commitment to strategy from organizational leadership.

Contributions to Research

Compared to other existing systematic analyses in the field, our theoretical advances consist of synthesizing and critically analyzing specialized literature to highlight advances, gaps, and convergences that had not yet been articulated.

The systematic review highlights that research on the application of AI in risk management tends to develop an integrated socio-technical theoretical framework, in which AI simultaneously functions as a tool, “decision agent and catalyst” of organizational resilience. Unlike previous reviews, this review offers a unified perspective on risk in integrated management systems, emphasizing the paradigm shift from control to adaptation, from predictive to explainable, and from technical approaches to ethical considerations.

4. Conclusions

The use of AI in IMS risk management is not merely a technological option but a necessity for organizations seeking to remain competitive, efficient, and resilient. The significant benefits of AI—from improving decisions, reducing costs, and optimizing processes, to increasing customer satisfaction—must be balanced by careful management of the ethical, technical, and organizational risks of using AI. AI serves as a transformative force for risk management in organizations with integrated quality–environment–OHSAS systems. It enables this shift by automating risk assessments, providing predictive analytics for processes and risk factors, and establishing a framework for continuous organizational learning.

The findings of this paper facilitate the formulation of strategic guidelines for integrating AI into risk management processes. These strategies encompass critical dimensions such as continuous monitoring for early detection, predictive and preventive analytics, the automation of risk assessments, and the synthesis of organizational feedback with IoT data to uncover latent risks. Furthermore, these proposed strategies present viable topics for subsequent research to evaluate risk management performance with IA within IMS.

The conclusion of this study reveals that the strategic integration of AI into the management system reduces risks through proactive detection, strengthening the resilience, sustainability, and safety of employees and the organization as a whole.

5. Limitations of the Study and Literature

The methodological scope of this study presents certain limitations. First, the reliance on six established academic databases (MDPI, Elsevier, SpringerLink, Emerald Insight, Taylor & Francis, and Google Scholar) and the exclusion of grey literature (e.g., industrial reports, unindexed conference papers) may result in a narrower thematic coverage. This is particularly pertinent for capturing state-of-the-art practical applications and developments in nascent fields. Second, with the literature search concluding in August 2025, findings published after this date are not reflected in the present analysis.

The existing literature on artificial intelligence in risk management is concentrated in sectors such as manufacturing, energy, and finance, with fields like construction and transportation remaining underrepresented. Furthermore, there is a notable scarcity of organizational case studies across all sectors. Addressing this gap through empirical case studies would provide significant value, enrich the scholarly discourse, and advance both theoretical and applied knowledge. Consequently, a key direction for future research involves the execution of cross-sectoral case studies to bridge this gap and foster a more comprehensive understanding.