3.1. Sustainability Management

A major current, global problem is sustainability. Public and private organizations across the world must become active agents for sustainable development. They must solve global issues (like poverty, hunger, or education for all) and plan targets for future generations. And little has been done in terms of global solutions to help these organizations to become sustainable and to manage this sustainability.

Current organizational sustainability requires new, revolutionary social and technological performance. It requires many changes in organizational targets, structures, principles, priorities, and values; it requires new strategies and new management solutions for the relationships that take place inside the organizations and with other organizations. Indeed, the new organizations must be socially responsible for future generations using new local and global partnerships and management.

Human resource management is vital for organizations and governments which intend to become sustainable. Activities and principles of this management must develop new talents needed for their globalization. However, “it is becoming increasingly more difficult to find competent, motivated workers who have good attitudes and work ethics” [4]. We believe that it is necessary to promote moral and ethical global solutions to stimulate, support, and manage staff development in superior frames of intelligence; in global, cooperative frames of cognition where individual development must be placed in lower abstractions and collective development in higher abstraction.

3.2. Education and Cooperation

Current global development of societies underlines the importance of education. Indeed, education is “one of the main drivers of the progress of a country” [5].

Nowadays, educational organizations in global environments require new collaborative actions for providing quality and efficient educational services to all. “The population does best if individuals cooperate” [6]; “cooperation is a key aspect of social evolution” [7]; “humans cooperate to build societies” [8]; “close cooperation between many entities for the benefit of public interest is one of the most important determinants for public sector organizations” [9]; “cooperation improves performance” [10]. Indeed, some educational systems are now capable of delivering some integrated services and collaborate for student satisfaction. These systems allow education professionals to make better system supported decisions for the learning of their students. For example, Moka and Refanidis [11] designed a system to automatically schedule the educational activities of a student. Using this, the student can take “informative decisions” to attend to learning objects. More recently, in [12], a collaborative online learning environment improves students in group’s attitudes involving social interactions.

The educational sector also needs a global environment where professionals and students can cooperate using global educational services. “Digital tools, online platforms, and collaborative technologies” are transforming the educational environments into a “globalized and interconnected domain” [13]. And for this end, we believe that it is important to design management mechanisms that create new global educational services and efficiently maintain these for our future generations. Indeed, sustainable education must preserve the quality and efficiency of appropriate global solutions across current organizations. And the support of government institutions and their management actions are “essential success factors,” promoting international cooperation around the world and increasing general awareness using social networks and media [14].

Further, the educational sector must contribute to the establishment of global citizens through the design of new studies dealing with social and ethical responsibility; new studies dealing with principles of “respect, equity, justice, honesty, and transparency” for human resource management.

3.3. Global Cooperation and Related Studies

Global cooperation is “difficult to achieve” [15]. Social, cultural, and language differences between humans and systems may cause major difficulties. And we cannot observe our biology to understand and support our true “cooperative nature.” Biological models, like neural networks, are not sufficient [15]. We need looking into human relations, global behavioral capacities, and moral thinking. We need numerous individuals globally acting together to produce relevant shared outcomes. Indeed, global cooperation requires that humans learn special, global capacities to construct positive collective actions. These abilities involve, for example, establishing shared objectives (like SDGs) and performing collaborative activities. Further, this cooperation requires specific regulations to appropriately coordinate all this positive global activity.

Cooperation has been studied by many researchers in Human-System Integration, Human-Machine System, and Human-Machine Teaming. For example, in [16], Hoc proposed the next definition for cooperation: “Human and machine are in a situation of cooperation if: (a) each of them strives toward goals although interfering with the goals of the other (at least based on resources or procedures), and (b) they try to manage such interference to make easier the activities of the other.” The first point highlights common objectives like SDGs. But this common goal could be the same for cooperation and for competition. Thus, we need appropriate frameworks to guarantee that our IT supports the positive objective of cooperation. We need management frameworks guiding the design of computational models able to support positive collective actions. However, more investigations will be required to connect with the positive nature of our actions (see also [2]).

3.4. Technologies and Services

“No nation, great or small, in the world today and tomorrow can secure its future alone” [17]. The global network of relationships is growing between individuals, organizations, and states. Companies and organizations at large produce new global services to increase prosperity for all citizens. And current technologies are important to facilitate this globalization. However, technologies can also cause “new spikes in inequality” [18].

“Today’s business is strongly influenced by globalization” [19]. “Managers and employees are exposed to global strategic decisions and cross-cultural interactions” [19]. Companies “operate in a complex environment” [20] characterized by the global economy and the integration of new IT based on current AI technology. They need “gathering, analyzing, and delivering” global information to help grow business and markets. And these global data are largely obtained using the web. However, the web needs more intelligence [21] and there are developing fields like “web intelligence” [22] to perform global services in the connected world.

The first organizations based on global services used “knowledge networks” built upon professionals managing “domains of knowledge” [23]. Since 2013, the world turned in an economy “driven by services” [24]. And “success in global service innovation required companies to develop capabilities that support increased relationship intensity and interactions” [25].

The knowledge management systems provide useful, global information to employees. They allow capturing, creating, and disseminating a “great deal of information” [26]. For example, Chioreanu et al. [27] designed an information monitoring platform for the industrial robot’s field. This architecture generally supports the management of industrial robotic systems based on IT services. It provides customizable services for cooperation among all the actors involved in the robotics sector.

3.5. AI and Management

More general management systems have also been implemented with recent AI technologies. For example, Qiaohong et al. [28] presented a middleware architecture for an intelligent management system. And this is an important basis for low-level software platforms. But intelligent management ideas need to be combined with current AI technology as an important part of the management practice [29]. In doing so, Information Communication technologies and recent Machine Intelligence techniques will be able to support new management behaviors. However, we must be careful that there is no complete replacement in this management.

Multiagent systems have been proposed for providing “computer supported cooperative environments” where “coordination, negotiation, and communication among various organization units” are required [31]. Agent-based models have also been used as computational models to “explore group behavior” in [32]. Tello-Leal et al. [33] also presented an “agent-based platform.” This allows “collaborations among heterogeneous and autonomous healthcare organizations focusing on the process-oriented integration.”

More generally, AI technology is currently used in public service organizations to improve policy decision making and to enhance public service delivery for citizens [34]. Further, AI applications are very successful for internal management issues [35] like cybersecurity or financial management. The existence of high-quality datasets as well as fast machine learning algorithms allow the successful completion of all these tasks with large amounts of data. However, recent studies like [36] underline the fact that there are also major risks in the AI technology adoption in governments.

This year, we have concluded some investigations on the design of intelligent software architecture for cooperative DLSs extensively used in governments and related organizations. With this research work summarized in [37], we have demonstrated how these systems require new architectural modeling techniques to incorporate new cooperative actions as well as new AI-driven services (mapped to concrete AI applications) implementing decision support and recommendation processes. Further, in [37], we have provided an evaluation case study for our software architecture to achieve SDG4, ensuring inclusive and equitable education worldwide.

3.6. ITIL Processes and AI

A process model shows “cause-effect relationships”; it shows “how the business entities are affected by the process actions” [38]. “A collaborative process defines the global view of the interactions between organizations to achieve common business goals” [33]. It focuses on “the coordination of activities to improve the management of resources and services” [33]. Process managers in the public sector need “cooperation skills to be able to build and to sustain relationships with citizens” [9]. However, this cooperation does not seem to be supported by current standards in process-oriented management.

IT Service Management (ITSM) has been used as a general framework to “implement and manage quality IT services” provided by “an appropriate mix of people, processes, and IT” [39]. It is concerned with the planning, organization, offering, deployment, and maintenance of IT services instead of IT technologies. More recently, ITSM has been adopted in combination with advancing AI technologies like, for example, chatbots in customer services [40], simulating the way humans would behave as conversational partners offering the IT services.

ITIL is a collection of best practices in ITSM; it provides a general framework for robust ITSM; it is a global process-oriented standard used in the development and governance of IT infrastructures. ITIL helps to create and maintain successful IT systems that “add value” to companies and organizations [27]. ITIL provides IT services with “better performance and quality” [41]. Further, ITIL helps to evaluate current knowledge built with social networking from the service management viewpoint [42].

With the advances of AI, the latest version of ITIL (ITIL v4) is being used in large environments with machine learning techniques. These investigations enable real-time decision making in complex markets. Further, they improve organizational responses by managing large data volumes (from the web) as currently required. However, the integration of ITIL v4 with current generative AI findings and with global, cooperation-oriented processes has not been clearly established and need to be investigated. Indeed, most ITIL tools support the work of individual actors rather than supporting the cooperation among “all the parties involved in the processes” [43]. With other words, “ITIL lacks cooperative perspectives” [44].

4.1. Intelligent Models and Abstractions

So far, we have presented three main types of models: “generic models”; “computational models”; and “operative models” (see Figure 1). As we describe in our recent publication [1], we consider these as intelligent models since they are mainly inspired by knowledge-based and behavior-based foundations of AI. Further, our generic models focus on cooperative intelligence, our computational models incorporate AI-driven services, and our operative models use concrete AI technologies like chatbots.

In this paper, we introduce a new type of model: the “management model.” As we shall see, the more we refine our generic models, the more we can achieve lower management abstractions to combine models of intelligence (not only computational intelligence) or adaptive models of intelligence. This refinement includes some concrete mechanisms (like learning, adaptation, or decision making) and/or procedures (like process activities) to really facilitate and support (never replace) management in future sustainable environments.

For example, a management model for a sustainable organization could be a globally evolved version of a knowledge management solution for an IT department. As a first global approximation, this type of management might be like a new computational model where the manager behaves as a kind of “stereotype” using knowledge. But this is not the global solution that we propose. We believe in “people and management” as a human solution based on global capacities; as a solution specially dedicated to improving thinking and behaviors; and as a set of open, natural procedures that help driving people (in their life, in their careers) toward sustainability.

Figure 1 illustrates our preliminary intelligent models and some relevant associations among them. In general, we can achieve computational models of intelligence from very generic abstractions based on organizational entities and their cooperative, manageable links. A more complete computational model supports generic abstractions of intelligence using concrete AI techniques and software engineering methods. And this computational model can facilitate the design of a subsequent management model. Indeed, we can design management solutions based on the support of computational intelligence. However, this is not required; we can achieve our management models without the support of any computation and/or architectural design. This last modeling technique represents a methodological solution. Further, we can use our management models to guide the lower-level design of our computational and/or operative models. Similarly, we can define operative models (very close to current technologies and standards) without management. We can integrate our technological solutions into computational models. The most appropriate design for a global organization might depend on the higher organizational entities and their associations.

In this paper, we organize and evolve some of these aspects. We investigate new global abstractions using three core designing elements: services, architectures, and processes. And a GS&POC model, as we describe later, summarizes some of these investigations. More generally, the paper brings five explicit abstractions to support intelligent cooperation from a globally oriented management perspective.

As we shall see, we also propose several refinement techniques for adapting (and evolving) each of these abstractions to our cooperative objectives, either locally (to improve local developments of intelligence) or globally (to solve the most helpful intelligence). All these abstractions are related throughout our general design methodology. We methodologically produce flexible and efficient abstractions across the organizations. And, following this methodology, it is possible, for example, to design very technical solutions for final operative models. Indeed, the last two abstractions that we present in this paper allocate a “methodological model” next to a “technological model.” The methodological model is mainly focused on the procedural part of the cooperative processes. The technological model is built upon the selection, distribution, and/or integration of currently available technologies like AI.

These refinement techniques introduce levels of detail to support superior, generic models of cooperative intelligence. Indeed, we support top, global cooperation in terms of collective system actions producing “global effects” like collective goals [45]. Further, these refinement techniques introduce levels of interaction, since cooperation arises from “interaction structures” [6].

4.2. Global AI

We investigate new computational solutions to support our intelligent models (including the management model) in the context of what we have called Global AI [2]. This new approach helps to improve human behaviors and thinking, supporting cooperative models of global, true, natural intelligence, as we describe in [2].

We work on the theoretical (not so much practical) integration of Knowledge-Based and Behavior-Based approaches to AI. And we have established new concepts, ideas, and general solutions to define this new, globally oriented approach to AI based on system cooperation.

In [2], we concentrate most of our efforts in the modeling of intelligent cooperation based on new global capacities, sustainable thinking, and moral behaviors. We model new integrated systems of humans and machines from a new, globally oriented perspective instead of just building the models from either the behavior approach (as initially introduced in [46]) or the physical symbol system (as widely applied in classical AI) or the most recent data-driven methods (like machine learning or deep learning) or the latest evolutions of knowledge approaches based on generative AI [47]. Then we design computational intelligence solutions to support this global view of intelligence. For example, in this paper, we present new, open management frameworks, architectural designs [37], and methodologies that can be adapted to any scenario such as a global, sustainable organization or a company, as we describe in Section 8. Further, we design some flexible and efficient low-level elements (e.g., global networks of system behaviors, knowledge registries, etc.) to computationally support our most generic, conceptual abstractions, giving new engineering insights into our Global AI approach.

The investigations that we present in [48] focus on the design of learning-oriented solutions for educational environments. This research work provides a complete architectural design for scientific behavior and knowledge, to convince people to share well-structured knowledge, and collaborating with each other in an intelligent, open, and globally efficient way. In practice, this type of investigations allows us exploring the design of new architectural solutions for intelligent eLearning systems between research centers, universities, government organizations, sociocultural entities, and industries. Further, we investigate intelligent software platforms [37] to be able to support our high-level abstractions of intelligent cooperation with physical and/or virtual models of computational intelligence.

In these learning-oriented designs, we suggest modeling generic eLearning environments and mapping these into computational interactions while adapting high level architectures. Further, we apply some classical AI principles to analyze generic eLearning environments as knowledge-based systems. And we exploit some behavior-based insights to design intelligent system actions for final, operative platforms. This largely hybrid approach helps us approximate workable integration among knowledge and system behavior from a computational (mainly virtualized) perspective.

Now, in this paper, we extend some of these initial ideas toward a more concrete foundation of our global approach to AI. Basically, we integrate what we believe in as the unique, global, superior intelligence [2] that surrounds us, with new abstractions and management techniques to support human behaviors and thinking. In essence, we explore the term intelligence and point to new balanced principles of ethical and moral development for people and organizations toward global cooperation.

More particularly, this paper presents the fundamentals of a general methodology to design a management framework for global, cooperative models of intelligence. We introduce an overall top-down approach to design IMFs in global environments. And, in this approach, we combine services, architectures, and processes, moving from high levels of cooperative intelligence to highly objective models of management and computational solutions to support these.

All these models and frameworks represent the result of several years of investigations in AI and innovative solutions for enterprise systems and IT management. Further, we have created all our abstractions and techniques following some of our initial ideas in behavior systems (like [3]), AI solutions (e.g., [2]), and intelligent software architectures (like [37]).

6.1. Generic Cooperation Abstraction

The first abstraction is a qualitative network of global organizations, a generic cooperative model among generic organizations and top-level cooperation. In this network, the nodes represent the organizations, and the edges show their cooperative and manageable links. It is a dynamic model for global organizations.

Figure 5 represents a global model with four generic organizational entities involved in global cooperation. These are Industrial Organizations (O1), Government Institutions (O2), Sociocultural Entities (O3), and Companies (O4).

Each of these organizational entities encapsulates some industry, government, sociocultural, or enterprise functions and objectives with cooperative values. It is a globally oriented institution that integrates moral/ethical behaviors and knowledge and computational models to support these. And in our approach, all the organizations respond to the same laws and ethical principles. The type of principles characterizing our global intelligence approach [2].

The cooperative links (edges) can be either directed or bi-directed arrows (see Figure 5). A bi-directed edge means that the organizations that it connects can cooperate (through knowledge acquisition and/or system behavior interactions) in both directions. A directed edge represents a single directed action and/or system behavior pattern.

In our computational models, every organization that they support represents either a private or a public institution with moral/ethical behaviors and thinking, as it corresponds to global intelligence. With generative AI approximations, however, the computational model would only support knowledge-based conversations with trained models.

The computational model of an organizational entity requires some kind of functional mechanism to support both knowledge acquisition and cooperative system behavior executions. It is an intelligent architecture that controls the registration and intelligent delivery of knowledge and the execution of moral, principled system actions to support top-level behaviors.

Each top-level organization can be divided into any suitable set of suborganizations and internal interactions. And this helps us to design more detailed collaborations inside and outside the organizations.

We define and use several types of interactions (e.g., collaborative tasks) to mean multiple, flexible low-level cooperation. All these interactions add new qualitative and quantitative information to our models. The overall map of interactions of a cooperative model describes the top cooperative model. The shape and complexity of this network depends on the decisions made during the design process.

The individual and organizational functions as well as the intelligent actions underline the introduction of computational intelligence in the approach. This is further detailed in the following abstractions.

6.2. Service-Oriented Abstraction

In our top-down approach, we propose a service-oriented refinement technique to create global, intelligent services at the second level of abstraction. The resulting intelligent services integrate knowledge and system behavior structures. They represent basic elements of knowledge and system behavior interactions.

The service-oriented refinement technique facilitates the evolution of our generic cooperative models into global platforms of intelligent services. Basically, this technique consists of analyzing global objectives and functions and describing these in the form of “services.” These service descriptions facilitate also the next process-oriented refinement, like in ITIL v4. Nevertheless, we can be flexible in the way that we approach the design of our GS&POC models.

For example, we can identify the architecture and related activities (basic processes) first and extract the global services at last. The only problem that we have found with this second approach is that it is difficult modeling the organizations without having any idea about the type of services that they can share through their cooperation. Further, we try to support high level intelligence and so we think that these sharable structures like services should be identified first, in terms of, for example, static forms of intelligence (e.g., system behavior patterns). Then we can design all the dynamic mechanisms and functions that help to control and process such system structures.

Using our approach, we define global services as basic combinations of single intelligent services, according to our higher-level abstractions. An intelligent service integrates knowledge and system behaviors. The final service model depends on the high-level network of interactions and the refinement process to support these. It represents a set of dynamic structures to help thinking and acting in global, cooperative networks.

In [37], we describe four major types of services: knowledge-based services, behavior-based services, hybrid services, and generative AI services, to mean that all these services can be mapped to special purpose AI applications based on classical or more recent AI techniques. For example, a knowledge-based service can be mapped to an AI application implementing decision making processes or recommendations based on traditional rules (e.g., fuzzy control rules) over the knowledge of the system (data and metadata registries describing knowledge). Similarly, a behavior-based service can be mapped to an AI application implementing decision making processes or recommendations based on trained models (using neural networks and learning algorithms) over the system behaviors and low-level actions of the system.

Furthermore, we refine our superior models in the form of services following a top-down technique largely inspired by the ITIL v4 framework. Following these standards, the global service strategies are identified first for making a unique environment of interactions. They provide direction (toward strategic objectives) and a vision for developing our global services. They are an integral input for global service design where most of our processes are involved during process-oriented refinement (see Section 6.4), including service catalog management, service availability management, and capacity management.

Global service development makes the services available for all their users, meeting the strategy and its objectives through processes like change management and deployment management. Finally, global service operations maintain the services to ensure that they operate as they were designed to. Some of the processes involved in these operations are incident management and problem management. But we believe that all these processes can be better approximated and methodologically defined once the architecture of the organization has been established.

6.3. Architecture-Oriented Refinement

The third abstraction represents an architectural approach to system design. In this level, we design, adapt, and/or integrate service-oriented architecture with more detailed knowledge-based and system behavior–based structures and engines. Nevertheless, we can also be flexible in the way that we add the architectural elements into our models and abstractions. Indeed, it is possible to identify and define new global services while we develop new computational models (with or without AI systems) to support human knowledge and system behavior networks.

In our previous work [48], we designed some registration and intelligent delivery engines (IDEs) for learning-oriented models. Here, we introduce new architecture designs to develop more thoughtful engines and system behavior modules in our abstractions.

Our intelligent architecture statically and dynamically generates global services using hybrid registries, service registration and management engines (RMEs), and service delivery engines to help higher level abstractions. All these functional elements serve to create, integrate, and control system behavior executions (like low-level system actions), knowledge acquisition, and/or decision-making processes using AI techniques.

A hybrid registry includes knowledge structures and system behavior patterns, as initially identified in service definitions (in the second abstraction). Intelligent engines compute higher level interactions (among mapped applications) according to this.

The RMEs create, store, and administer intelligent services. The delivery engines process and intelligently deliver these services. These engines implement intelligent processing mechanisms to support top-level cooperative tasks. Its underlying control mechanism highlights relevant “AI” being processed by the engine. Further, behavior-oriented AI techniques can be used to train new structures so that architecture can be observed to learn new system behaviors (and lower-level actions).

We use these registries and engines as flexible design elements for our computational models. Indeed, we can integrate these functional elements at any abstraction of an intelligent cooperation. For example, Figure 6 shows a top computational model based on intelligent services registries, RMEs, and IDEs. In this computational model, some industrial and government organizations can register and deliver intelligent services using those engines. The employees of the industrial organization can internally register and manage some services while publishing other services into the government organizations. Similarly, the employees of the government organization can register and deliver globally oriented services into the interconnected (cooperatively linked) industrial organization.

When we integrate registration, management, and delivery engines into a computational model, we make the underlying mechanisms and systems processing the services of the modeled cooperation. In essence, what this cooperative network means is that we can support a generic cooperative model with a functional architecture where all the participants can access and execute global services. Every time an employee makes a request through a delivery engine, the mechanisms and controllers of this functional module collect the corresponding service templates and execute knowledge processing and system behavior interactions accordingly. Thus, the resulting architecture is “almost operative.” However, it requires some method to describe how to register, manage, and deliver these services. And this is what we define in the following process-oriented refinement technique.

Using current AI techniques, architecture-oriented refinement would take advantage of AI applications based on data-driven computational models trained with large and changing datasets (using neural networks and learning algorithms). This would reduce the computational efforts in developing the applications mapped from the services at the cost of energy in applying the learning methods.

6.4. Process-Oriented Refinement

The fourth abstraction is based on process orientation. Here we introduce methodological aspects of design to continue supporting our top-level, intelligent cooperation. We identify and integrate lower architectural elements of development with some general procedures to complete a general, abstract definition of IMFs based on some ITIL v4 processes.

An ITIL process is a “structured set of activities designed to accomplish a specific objective” [39]. And every process needs management to control its overall performance toward efficiency and objective achievement. These control mechanisms and the process enablers (resources and capabilities) form the process model. The most important resources of the process model are people, capital, infrastructure, application, and information. Process managers ensure that the processes are executed correctly.

In our approach, we define a global process as a set of roles, responsibilities, and activities that encapsulate the main features of our higher-level abstractions. One global process can include one to many of these process elements. The roles and responsibilities define who is responsible for doing what, in terms of creating and/or managing the global (and local) services and functional engines of our architecture. The activities formally describe the methods for these processes.

The resulting processes around these services establish most of the methods (activities) that are required to convert our computational models (with or without architectural elements and/or AI systems) into operative models of intelligent services. These models behave as highly distributed systems. They incorporate hybrid intelligent mechanisms and processes to execute the services. The final users of these operative abstractions contribute to their global networks, either creating new cooperative knowledge or executing intelligent system actions according to the global and/or local processes. Thus, a final user can be either a human, an automated service, or another basic process.

Using our global processes and high-level architecture, the employees of an organization could search, find, and/or create new services based on, for example, individual and/or organizational system behaviors. The managers can supervise and authorize these new services, according to our process-oriented abstractions.

ITIL processes can be used in combination with AI techniques. For example, Zarrazvand and Shojafar [41] used fuzzy cognitive maps to model “the problem of needing ITIL processes.” In [40], chatbots are used to simulate how humans behave as conversational partners offering the IT services. In our top-down approach, we can also introduce AI techniques and, more particularly, learning methods to drive data among lower-level elements of our architectures. Further, we can incorporate online chatbots to support intelligent language models and conversations for user-oriented services.

6.5. Technology Abstraction

Finally, we propose a technical abstraction (the fifth level) based on more detailed architectural orientation and processes. In this last abstraction, we introduce some technological aspects of development to continue supporting our cooperation. More particularly, we create, adapt, and evolve special purpose technologies to complete our GS&POC models. And we design final IMFs using efficient technologies like, for example, AI data-driven methods. But we can be flexible and efficient in the way that we adjust our technologies and related standards.

The employees of an organizational entity can specialize in any of the available abstractions. For example, a software engineer could specialize in architecture-oriented abstraction for intelligent SOA constructions. Model builders (and other related roles) can construct system abstractions from higher level designs, looking at local components (e.g., in one organization) or global elements (e.g., knowledge sharing interactions). Further, web technologies and web intelligence techniques can be incorporated at this level of abstraction to develop the knowledge registries and their interfaces. Further, we can refine our architectural components at the level of data so that new AI data-driven methods can also be incorporated to automate part of the higher-level models.