1 Introduction

“Human-centricity” has emerged to become a characterizing feature of the next -stage of Industry 4.0 [1]. One way to describe the rise of human-centricity in manufacturing is the progression from focusing on reducing the operator’s physical labor to reducing the operator’s mental labor [1], 2, and the shift from prioritizing shareholders to stakeholders [3], and the rise of value-driven over technology-driven manufacturing [4]. Two technological advancements play key roles in human-centric manufacturing: Artificial Intelligence and the Industrial Metaverse.

The term “Intelligentization” first came to existence by the Chinese State Council in its National Artificial Intelligence Strategy that details milestones until 2030 [5] and translated as “Next Generation Artificial Intelligence Development Plan”. The use of Artificial Intelligence (AI)-based technologies has become the main characterizing feature of Industry 4.0. (Fig. 1). The term Industry 4.0 (spelled Industrie 4.0 in German) became widely used due to its unique definition that summarizes the evolution of industry.

Fig. 1
figure 1

The four generations of Industry and the characterizing feature of each: 1.0 (Mechanization), 2.0 (Automation), 3.0 (Computerization, and 4.0 (Intelligentization).

Full size image

On the other side, the term Metaverse came to a widespread use after Facebook, Inc. rebranded itself as Meta in October 2021 [6], resulting in an explosive use of the term on the World Wide Web. In December 2021, the Chinese tech giant Baidu launched its metaverse platform 希壤, translated as “Land of Hope”. TikTok followed suit in early 2022 with its Metaverse platform 派对岛, translated as “Party Island”.

In the work of Chang et al. [7] an emphasis on the integration of 6G-enabled edge AI with the Metaverse is shown, exploring architectures, challenges, and future directions specific to this technology. They specifically address the technicalities and applications of 6G technology in enhancing Metaverse experiences, like immersive education, smart cities, and product testing.

A discussion of the implementation of flexible and reconfigurable smart factories within the framework of Industrie 4.0 was shown in [8], in which emerging technologies such as the Internet of Things (IoT), wireless sensor networks, big data, and cloud computing were integrated into the manufacturing environment. It presents a detailed conceptual design of smart factories, exploring their operational mechanisms and technical features, and outlines the benefits of smart factories, such as flexibility, productivity, resource and energy efficiency, and transparency.

The emerging concept of MoT, which blends IoT, AI, and blockchain within the context of 5G and 6G networks to enhance metaverse and XR applications is presented in [9], with an emphasis on decentralized service management and data flow synchronization in 6G networks.

In the work of Xu et al. [10], the technological challenges for implementing the Metaverse were discussed, such as communication, networking, and computation perspectives, including the use of blockchain for interoperability and resource management.

In this work, we focuses on the evolution of Industry 4.0 through the integration of AI and the Industrial Metaverse, emphasizing human-centric manufacturing, intelligent industrial robotics, UAVs, and additive manufacturing. Technological scope and application presented in our manuscript is more broad than in [7, 8, 9, 10]. In this work, we cover various aspects of Industry 4.0 technologies, including detailed discussions on robotics, UAVs, additive manufacturing, and their challenges. In addition, we highlight distinct challenges here such as safety concerns and workforce skill gaps in Industry 4.0 technologies. In the aforementioned works, challenges such as privacy, network latency, and resource allocation in the context of the Metaverse were the focal challenges.

Although there is no uniform definition of the Industrial Metaverse, the most acceptable definition is that formulated by the Massachusetts Institute of Technology in its periodical MIT Technology Review in partnership with Siemens, which defines the Industrial Metaverse as the “Metaverse sector that mirrors and simulates real machines, factories, cities, transportation networks, and other highly complex systems—and will offer to its participants fully immersive, real-time, interactive, persistent, and synchronous representations and simulations of the real world” [11].

Thus, to understand the definition of the Industrial Metaverse, it is imperative to define the term Metaverse. According to the Encyclopedia Britannica, Metaverse is “a proposed network of immersive online worlds experienced typically through virtual reality or augmented reality in which users would interact with each other.” It should be noted that the term is continuously evolving, thence the absence of a uniform definition [12].

Therefore, under this definition, the industrial metaverse integrates several technologies that include among others digital twins, blockchains, and cloud computing.

However, during the COVID-19 (Corona Virus pandemic of 2019), the world witnessed severe supply chain disruptions that crippled every manufacturing sector and paralyzed the transportation sector (particularly air transportation). The pandemic exposed a critical need to develop local, regional, and national self-sufficiency and an urgent necessity to develop plans to counter similar future disruptions in vital tasks that require physical presence [13]. The pandemic and the adaptations it engendered have resulted in an accelerated development of the technologies that aid in contactless-manufacturing, and that can counter supply chain disruptions, such as intelligent industrial robotics, unmanned aerial vehicles (UAVs), AI-empowered Additive Manufacturing, Digital Twins, Immersive and Augmented Reality. These technologies can be grouped under the two technological advancements under investigation: Artificial intelligence and the “Industrial Metaverse” [14].

2 Scope and methodology

In this work we performed a holistic literature review to answer the research questions: What is the main feature that characterizes Industry 4.0? and what are the associated technologies? and what are the challenges that are faced by these technologies? Particualrly, a light is shed on the safety aspect of these technologies in this paper. As was pointed earlier, human-centricity stands out to be the ultimate objective in Industry 4.0, which is also referred to as Society 5.0 in Japan [15]. The technologies that shaped Industry 4.0 can be grouped under two technological advancements: Intelligentization of industry, and the Industrial Metaverse. The thumbprint of these two technological advancements can be thought of to be similar in impact to Mechanization on Industry 1.0 or Automation on Industry 2.0 or Computerized Automation on Industry 3.0 (See Fig. 1). The impact of intelligentization will only be investigated through three technologies that emerged strongly in the last decade, particularly during the CoVid-19 pandemic: Intelligent industrial robotics, Unmanned Aerial Vehicles, and Additive Manufacturing. The expected challenges in each of the respective technologies aforementioned will also be investigated.

3 Intelligentization in industy

3.1 Overview of challenges

The fast deployment of AI-based technologies in almost every industry and every application has primarily created a significant demand and resulted in a severe shortage of a skilled AI workforce on two levels: the applied technology-handling level and on the research level. This shortage has been rapidly widening and is starting to have a tangible impact [16].

A study conducted by the Korn Ferry Institute [17] found that by 2030 there will be a global shortage of 85 million skilled personnel and that the shortage of human talent will be the biggest challenge in the field by that year. On the research level, the situation is much more urgent. This is because in-demand AI-related researchers require long years of advanced education and cannot be remedied by short training programs. For a country like the United States which still relies heavily on foreign talent, attracting that talent when the world-wide AI workforce is already depleted will become increasingly difficult.

These studies are supported by other research studies that conclude that the skill gap will be significantly wider than was previously expected. An Accenture survey found that 84% of manufacturing executives indicated that the workforce is unprepared to adopt the coming technologies [18, 19, 20, 21]. Previous studies conducted in 2015 indicated a shortage of 2 million jobs in advanced manufacturing in the next decade due to the skill gap [22, 23], which is already a large number to cover.

The preparedness for the tsunami of skilled workforce shortage has proved to be fruitful for countries that had invested considerably over the last decades in preparing the workforce. Taking China as an example, in the year 2010, for the first time in modern history, China surpassed the United States as the world’s largest manufacturer [24], and by the end of 2021, it had almost double the value-added manufacturing in the U.S. as seen in the most recent data by the World Bank (see Table 1) [25].

Table 1 Industrial ranking of the top 10 countries by value-added manufacturing, World Bank, 2021 Data [26]
Full size table

As pointed earlier, in the next sub-sections, we will investigate the recent advances in applications of the three AI-based technologies that played a crucial role in the last few years, particularly during the CoVid-19 pandemic: Intelligent Industrial Robotics, UAVs, and AI-empowered AM.

3.2 Intelligent industrial robotics

Table 1 Before the Covid-19 (Corona Virus of 2019) pandemic, the forecasted market size for the Intelligent Industrial Robotics market was $14.29 billion in 2023 from $4.94 billion in 2018 [27, 28]. In its 2021 report, the International Federation of Robotics (IFR), states that there were more than 3 million industrial robotic units in the world in 2020. Table 2 lists the top 10 countries in growth of industrial robotics for the years 2016, 2019, and 2021. The data shows that the East Asian markets (China, Japan, Korea, and Taiwan) alone constituted 71% of the top 10 market size in 2016, and steadily grew to 77% in 2021.

Table 2 The top 10 markets of industrial robotics
Full size table

Furthermore, the need for industrial robotics sharply increased with the Covid-19 pandemic, due to the added demand for no-contact material handling in applications beyond the manufacturing industry [32]. One of the major impacts of the pandemic was the exposure of the urgent need for more preparedness in developing self-sufficiency in the supply chain [13, 32, 33]. As a result, the supply chain disruptions seen during the pandemic have abruptly accelerated the use of next-generation intelligent robotics [28, 34]. The impact of the pandemic on the automotive industry caused it to lose its place as the largest robotics market to the electronics industry, as seen in Table 3.

Table 3 Industrial robotics market by industry
Full size table

The technical specifications of six collaborative robots are shown in Table 4, and technical data of traditional robots that are used for similar applications are shown in Table 5. The Fanuc CR-35iA collaborative robot is currently the collaborative robot with the highest payload capacity in the market of 35 kg, and the Swifti robot by ABB has the highest maximum TCP speed of 5000 mm/s. The Yumi by ABB has two robotic arms with a total of 14 axes, while the light weight Denso Cobotta demonstrates the super flexibility collaborative robots can achieve. Two commonly used collaborative robots UR3 by Universal and LBR iiwa 7 by Kuka Robotics are also listed to highlight differences with traditional robots.

Table 4 Technical specifications of collaborative robots
Full size table
Table 5 Technical specifications of traditional robots
Full size table

When compared to traditional robots, the payload capacities and maximum speeds of collaborative robots are far behind. TCP speeds of 11,000 mm/s are not uncommon for traditional robots.

Pictures of collaborative and traditional robots are shown in Fig. 2. It can be inferred from Table 3 and 4 that the primary technical drawbacks of collaborative robots compared to traditional robots lie in two aspects:

  1. 1.

    The Tool Center Point Speed is considerably lower in collaborative robots than in traditional robots.

  2. 2.

    The payload capacity is considerably lower in collaborative robots than in traditional robots.

Fig. 2
figure 2

Collaborative robots: The Denso Cobotta (left), Rethink Robotics Sawyer (left-center) [46], Traditional Robots: Fanuc M-iA LR Delta Robot (right-center), Denso VS087 (Right)

Full size image

The significantly lower speeds and payload capacities are a direct result of the safety concerns due to the nature of operations in which the collaborative robots are expected to operate. As is discussed later in this section, the safety standard ISO/TS 15066 limits the momentum of collaborative robots to reduce risk of injury in collisions. As research progresses and more data on safety in operation is collected, larger industrial collaborative robots can safely work with operators [41].

Aside from the much lower speeds and payload capacities, the safety concern remains the primary factor that contributes to the under-deployment of collaborative robots in industrial settings [42, 43].

Despite that, the International Federation of Robotics forecasts that collaborative robots will be leading the robotics industry in the upcoming years [44]. During the Covid-19 pandemic, collaborative robots were effectively used to support the implementation of social distancing in factories [45].

The primary operational advantage of collaborative robots over traditional robots lies in the significant flexibility achieved through a much less-restrictive standardized safety functionality. For traditional robots, safety standard ISO 10218–2 requires safeguarding for operator safety using physical barriers or other means such as electro-sensitive protective equipment that are compliant with the standards IEC 61496–1 and IEC 62046 [47, 48, 49]. In contrast to that, the principal standard related to collaborative robots is the ISO/TS 15066 standard, which was created to sophisticate the requirements in ISO 10218–1 and 10,218–2 in regards to collaborative robot speed and distance to ensure safety, without the physical safeguarding. The safety functionality in collaborative robots is mainly achieved through Speed and Separation Monitoring (SSM) and restricting the momentum of the robot such that any collision with an operator will not result in an injury. SSM in collaborative robots is governed by the minimum protective distance, S, at time t0, which is modeled in in the standard ISO/TS 15066 by the following equation [50, 47]:

$$S\left({t}_{0}\right)\ge \left({\int }_{\tau ={t}_{0}}^{\tau ={t}_{0}+{T}_{R}+{T}_{S}}{v}_{H}\left(\tau \right)d\tau \right)+\left({\int }_{\tau ={t}_{0}}^{\tau ={t}_{0}+{T}_{R}}{v}_{R}\left(\tau \right)d\tau \right)+\left({\int }_{\tau ={t}_{0}+{t}_{R}}^{\tau ={t}_{0}+{T}_{R}+{T}_{S}}{v}_{S}\left(\tau \right)d\tau \right)+(C+{Z}_{S}+{Z}_{R})$$
(1)

The equation is an updated version of the previously used linear version of the equation:

$$S={v}_{H}\left({T}_{R}+{T}_{S}\right)+{v}_{R}{T}_{R}+B+C+{Z}_{S}+{Z}_{R}$$

where:

\({v}_{H}\) is the speed at which the operator is traveling toward the robot,

\({v}_{R}\) is the speed at which the robot is traveling towards the operator,

\({v}_{S}\) is the speed of the robot while braking.

\({T}_{R}\) is the response time, or the time it takes for the robot to react to the presence of a operator,

TS is the time the robot takes to come to a controlled stop.

B is the distance traveled during braking.

C, ZR, and ZS represent the intrusion distance safety margin based on expected human reach, robot position uncertainty, and sensor uncertainty, respectively.

In 2020, the standard update ANSI/RIA R15.08–1-2020 introduced formal definitions of the terms Autonomous Mobile Robots (AMRs), and referred to them as Industrial Mobile Robots (IMRs). The standard detailed the safety requirements for IMRs. Before the standard update R15.08–1-2020, there was no safety standard for mobile robots nor any official definition for them, other than a “Guidance for Manufacturers” that was published by the RIA under R15.08.

In addition to these standards, RIA TR15.806–2018 detailed the testing procedure for the power force limiting systems found on most collaborative robots. Table 6 shows the main standards related to collaborative robots.

Table 6 Safety standards related to intelligent industrial robotics
Full size table

3.3 Intelligent unmanned aerial vehicles

Unmanned Aerial Vehicles (UAVs) or more commonly known as drones have effectively been used in Search and Rescue (SAR) and disaster relief operations in the aftermath of earthquakes, hurricanes, explosions, and other disasters [51, 52, 53]. UAVs have also been used in mapping and observations [54, 55], inspecting buildings and bridges [56, 55], logistics [57, 55] construction management [58, 55], and exploring harsh environments that are unfit for humans [59, 60], and in what came to be known as “Smart Farming” or “Farming 4.0” for environment monitoring, image acquisition, and even irrigation and fertilization [61].

During the CoVid-19 pandemic, drones equipped with infrared thermal imaging cameras were used to locate individuals with fever symptoms, in addition, cameras connected to Artificial Intelligence (AI)-enabled systems were used to aid in locating individuals via facial recognition [62].

Recent advances in visually-aided navigation, high-precision sensory and perception systems, and highly-sophisticated attitude controls, allowed drones to be safely used indoors [63], such as part and material delivery in manufacturing plants [64], detecting problems in manufacturing equipment [65], conducting routine inspections in areas that are difficult to reach [59], detecting gas leaks, overheating machinery [66], and fire [59, 65], and in surveillance of manufacturing facilities [59].

The use of Artificial Intelligence (AI)-based algorithms in drone navigation can decisively contribute to their increased use indoors, by object identification, path planning, and determining traffic movement in order to safely navigate through a manufacturing plant, in which there is a significant mobility of machines and personnel [67].

Furthermore, the use of the next-generation of AI, known as “Swarm Intelligence (SI)” can significantly revolutionize the use of drones in everyday applications. In which swarms of robotic drones can collaboratively work together to achieve tasks. SI is inspired by insects such as bees and ants that naturally coordinate to accomplish tasks that otherwise would not be possible to accomplish alone [68].

The intelligentization of the drone industry has rapidly accelerated the growth of the unmanned aerial vehicles (UAV) market, which is expected to peak at $52.3 billion in 2025 from its $20 billion market size in 2018 [27, 60].

3.4 Intelligentization of additive manufacturing(AM)

Additive manufacturing grew quickly from only rapid prototyping to adding a high level of flexibility to manufacturing processes [69, 70], part repair, reduced energy consumption and manufacturing times [71], direct manufacturing of parts [72, 73, 74], and manufacturing of highly complex parts that otherwise require high costs with conventional methods [75].

However, recent research suggests that several challenges still exist in the field, in particular, low accuracy when compared to subtractive manufacturing techniques [76], and surface roughness of the final product continues to be a key challenge in the field [77, 78]. In addition to that, the lack of established and standardized methods of quality assurance of additively manufactured parts, makes this technology slowly progressing in the manufacturing industry, in which failure of components is minimally accepted [79]. Up until recently, the lack of optical transparency was another major drawback of additive manufacturing, in the work by Kotz et al. it was shown that highly-transparent silica glass was 3D printed with resolutions down to tens of micrometers [80].

Despite these challenges, AI-empowered 3D printing can provide a potential remedy to smart biomedical implants and directly prepare 3D objects on the target surfaces of the patients [81]. As AI is integrated in AM, and with the increased accessibility of AM to everyone, in-situ printing will be possible and even, it would be possible to fabricate a biomedical device in a short time on the patient without need for waiting an ambulance and delaying critical care [82]. In their work, Alex et al. used Artificial Intelligence (AI) assisted method to solve optimization problems, in which they used predictor-evaluator network (PEN) approach to train an Artificial Neural Network (ANN) and generated topology-optimized geometries. The resulting geometries are identical to conventional topology optimizers with an advantage of less computational power and time. In addition, the selected method does not need any optimized data to train the ANN [83]. In their work, Zhang et al. demonstrated the development of an AI system to repair the damaged part in 3D printing, particularly electrical wiring and structural system. A camera detects the damaged part and sends the information to the semantic segmentation and DRL module where grooves and circuit disconnections are repaired through composite printing. The reported system efficiency that was achieved was almost three times more than the conventional methods, the approach demonstrated how the integration of AI in the AM process allowed for handling of tasks in complex changeable environments [84].

Schneller et al. report a method on using applied AI to predict the fatigue failure behavior of the 3D printed parts. The input parameters that were used include the load stress, hardness and defect size. The parameters were trained on ANN and a binary output was generated as either failure or not. K-fold cross validation was used to tune the hyperparameters and architecture of the neural network. The accuracy of more than 91% was achieved for the prediction of the material fatigue failure behavior. The technique demonstrated how the expensive material testing and the different error-prone assessment methodologies could be replaced with a quick and reliable AI methodology [85]. Jin et al. applied AI using real-time camera images and deep learning algorithms to prevent the occurrence of the delamination and warping in the printed parts. The developed algorithm can measure the amount of warping, estimate the start of the distortion and extent of the deformation before it happens in the printed part [86].

4 The industrial metaverse

Recent research works suggest that the Industrial Metaverse will demarcate the next-generation of Industry 4.0 (also commonly referred to as Industry 5.0) [87, 12]. The technologies that it encompasses include Digital Twins, Augmented Reality (such as Projection and Location-Based Augment Reality), Virtual Reality, Mixed Reality, and Blockchains [88]. The expected impact of the Industrial Metaverse is very significant that it can effectively revolutionize the manufacturing industry [14]. For example, when virtual “digital twins” of industrial systems are integrated in the metaverse space, corporate managers can have real-time access to every plant they operate throughout the world, with their avatars walking the 3D virtual floor of the plant and interacting with the avatars of the plant managers. As another example, the blockchain allows for reliable financial transactions for traders at reduced costs.

Despite, the great potential of the technology, the Industrial Metaverse is still in the early stages. A comprehensive review of the integration of blockchains in the metaverse is conducted in the works of Mourtzis, Angelopoulos, and Panopoulos [14] and Yang et.al [89], and show that a wide number of applications can potentially setup an economic environment in the metaverse that includes among others Non-Fungible Token (NFT) markets, social networking, and businesses.

Despite the great potential of the immersive internet created by the Metaverse, there remain many challenges that can slow down its progress in the near future. These challenges include:

  • Lack of Standardization and Regulation As the technology is still in an early stage, standards and regulations related to the technology would also be in the development stage. In their work, Dohler et al. [90] show that three regulatory challenges standout in the metaverse: regulations on privacy, copy right laws and regulations, and enforcements of these regulations. When blockchains are used for transactions, all transaction histories are visible to public. Even if cryptographic hashes are used to hide transactions, electronic addresses of the wallets and assets are visible, consequently privacy regulations such as the California Consumer Privacy Act (CCPA) cannot be supported [90, 14].Another challenge arizes in the potential abuse of technology; immediately after the Horizon Worlds platform was released by Meta Platforms, Inc harassment incidents were reported [91]. Research shows that the trauma caused by assault and sexual harassment in the immersive environment can result in psychological, physical, and social impacts on the users such as increased heart rate and other anxiety measures [92]

  • Cybersecurity Concerns This is the primary and the decisive challenge that limits the use of the industrial metaverse as was shown in the works of Mourtzis et al. [14] and Yang et al. [89]. They identify a need to establish privacy and security structures in the metaverse as well as suitable cryptography mechanisms. In the metaverse environment a large amount of data is acquired from the users that not only pertain to financial transactions, but also include biometric data, user behavior [93] and hand gestures collected by the AR/VR/XR devises [94].

  • Cost the equipment needed for the technology as well as the computational power needed can limit the implementation of the technology. When blockchains are used for transactions, the operating costs include significant energy consumption, for example, Bitcoin mining consumes 60–125 TeraWatt-hours annually, which is equivalent to the annual consumption of Norway [95]. Although operating costs can be reduced due to the use of decentralized networks when compared to the conventional transaction systems that require a central authority, however, blockchain energy consumption remains a challenge [14].

Current generations of augmented reality and virtual reality have demonstrated the potential for supporting significant improvements in industry, particularly in maintenance and quality control. For example, Lockheed Martin reports that the use of Augmented Reality (AR) has resulted in a 96% improvement in accuracy in the inspection of its F-22 and F-35 fighter jets [96]. AR Smart Glasses were used by Boeing to guide electricians in wiring [97], in a similar way, projection AR is used to guide workers in the automotive assembly plants. Augmented Reality has reduced the necessity for physical presence for predictive and corrective maintenance procedures in manufacturing industrial plants. On the other end, Virtual Reality (VR) has been effectively used for training and predictive maintenance procedures.

When Augmented Reality is integrated in the AM processes, significant improvements in speed and optimization can be achieved. In their work, Cai et al. changed the configuration of the Material Extrusion technology by using digital twinning, and tested the results in Augmented Reality. The authors developed a reconfigurable AM robotic arm and its digital twin which was used for planning and simulating the toolpath. AR was used to provide the communication between the twin robots which resulted in an increase in the speed of the 3D printing process and promoted the exchange of the optimized layout in the digital twin and the physical system [98]. In the work by Yu and Xu, digital twins, augmented reality and cloud manufacturing were combined with AM and applied to human–machine interactions, which allowed for facilitating communications [99]. In another study AR was applied to AM processes to easily adapt to this technology and learn existing Design Heuristics for AR. For the proposed aim, the authors used the Design for AM cards with modification, which allowed the final printed part to be visually inspected before it is ready. Evaluations were also carried out to see the advantage of the AR application [100].

5 Results and discussion

Two technological advancements characterize the human-centric Industry 4.0, Artificial Intelligence and the Industrial Metaverse.

Both AI and the Industrial Metaverse share the challenge of the lack of the skilled workforce that is trained to handle the advances in the technologies. Almost all research studies agree that the lack of the skilled workforce will be the primary challenge for the realization of artificial intelligence and the industrial metaverse in manufacturing.

While the industrial robotics market grew rapidly, the share of the intelligent industrial robots in the market is still below expectations [42, 43]. This can be attributed to three factors: the first one and the main reason for the under-deployment is the safety concern that manufacturers have on depending on the sensory and perception systems that intelligent robots are equipped with for operator safety. A second reason is the reduced momentum that collaborative robots have compared to traditional robots as evident in the much lower Tool Center Point (TCP) speeds and payloads. Finally, the lack of qualified personnel that are trained on handling the technologies that collaborative robots are equipped with plays a decisive role in the deployment of collaborative robots. Consequently, manufacturers still prefer the use of safeguarded high-speed high-payload traditional robotic systems for which qualified technicians are readily available, over entrusting the sensory and perception systems on the low-speed and low-payload collaborative robots.

For UAVs, the primary challenge lies in the safety concerns on the use of the drone technology indoors, and the absence of clarity on legal issues that may arise from the use of drones outdoors.

The primary challenge for incorporating additive manufacturing in manufacturing is in the production yield and the lack of established and standardized methods of quality assurance of additively manufactured parts. While subtractive manufacturing methods may be more complex and expensive, and less flexible than additive manufacturing, they can give very high production yields, and there are established quality assurance procedures at every stage in the manufacturing process. In addition to that, surface roughness of the final product is a key challenge for the incorporation of the technology in some fields such as in optical systems.

The industrial metaverse is still very novel, many researchers agree that it will have a significant impact on manufacturing and other fields, as evident in the massive investments in the field. However, the major concern for incorporating the metaverse lies in the cybersecurity concerns and lack of regulations, in addition to the lack of the skilled workforce.

The major challenges for each of these technologies are listed in Table 7.

Table 7 Challenges in the adoption of the three technologies in manufacturing
Full size table

6 Conclusion and future perspectives

In this work we investigated the recent advances in Industry 4.0. Two technological advancements that characterized the human-centricity in manufacturing were under investigation: Artificial Intelligence (AI) and the Industrial Metaverse. In regards to AI, three AI-enabled technologies that played a crucial role in the last few years were under the spot light:

  • Intelligent industrial robotics,

  • Intelligentization in additive manufacturing

  • Intelligent UAVs.

Intelligent industrial robotics remain a corner stone in the manufacturing industry. Research studies show that the safety concern is a key issue that contributes to the below-expectations of the adoption of the technology. Safety and legal concerns play a major role for incorporating UAVs. For additive manufacturing, quality assurance and production yield concerns remain main challenges, and cybersecurity is the challenge that will need to be addressed for the industrial metaverse.

Almost all researchers agree that the lack of the skilled workforce will be the primary challenge for all Industry 4.0 technologies in the next decade.