1 Introduction

In the upcoming 10–15 years, the growing adaption of autonomy and teleoperation will transform the workplace, with an ever-increasing human–machine interaction. The utilization of information technology (IT) and artificial intelligence (AI) has been boosting productivity and adding pivotal value for companies across sectors from manufacturing, education, and media to transportation or customer service. Having been a primary driving force for big data, robotics, or the Internet of Things (IoT), artificial intelligence leverages high-computational power to harness massive amounts of data by using trained models through machine learning or deep learning algorithms [1, 2]. This ongoing shift has been transforming and reshaping the required skills of workers inevitably, as was the case since the first industrial revolution, creating skill shortages or mismatches in the labor market with a consequential negative impact on the economy [2].

With the introduction of the concept of Industrie 4.0 (Industry 4.0) in 2013 [3], powerful, autonomous microcomputers (embedded systems) have been wirelessly networked with each other and with the Internet, that results in the convergence of the physical world and the virtual world (cyberspace) in the form of Cyber-Physical Systems (CPS) [3]. The impact of the ability to “network resources, information, objects, and people to create the Internet of Things and Services” is expected to inevitably reshape the industry that brings the concept of the fourth stage of industrialization.

Romero et al. [4] state that the paradigm shift by Industry 4.0 has been changing the organizational ecosystems and the roles of the workforce, embracing state-of-the-art technology, thus, the overall expectancy from the human capital. The asset of human capital is defined as the combination of skills, education, and expertise that form a base for well-being and productivity in the work environment. The Industry 4.0 organizational hierarchy is decentralized, leading to more coordination, interaction, and communication that ultimately requires a creative and strategic workforce that is cross-functional in performing/operating activities or processes. These factors centralize the human factors even more, as human capital is the leading component of this paradigm shift. This interconnected, dynamic, and collaborative nature facilitates its integration into business ecosystems that will influence varying sectors to a greater extent inevitably [4].

The key drivers of future work can be grouped as technological shift, generational shift, and career shift, where the industry needs systems that facilitate skill engagement through life time learning, upskilling their existing workforce, and providing grounds for workers to transfer their skills to new jobs [5]. Roughly speaking, automation and technological improvements alter the nature of the jobs and associated tasks, decrease the number of routine jobs in the near future, and introduce new professions that involve new sets of skills. Ultimately, in developed countries workforce might be threatened because of the higher cost of labor; however, the economic impacts are less critical due to the diversity in the overall economy. For developing countries, however, the shortages in skills might motivate technological solutions, and the economic impacts are likely to be more prominent if the country’s economy is dependent on the evolving sector [5].

The mining industry today faces with several technological advances, including digitalization and digital transformation in several aspects of the production line, introducing challenges of cultural changes and adaption in the mine environment. Smart mining, intelligent mining, or Mining 4.0 (meaning Industry 4.0 in mining) are some terminology commonly used for this evolution that refers the adaption of digitalization in numerous operational fıelds in the mining industry, leading the transformation in business models. In order to identify the digital technologies adopted in the mining industry so far, Barnewold and Lottermoser implement text mining analysis along with network analysis to establish the relations between identified technologies, utilizing insight reports of the mining industry, published scientific papers, and media data. After identifying the relevant digital technologies in the mining sector, the researchers establish a relationship between these digitalization technologies and their adaption/implementation in active mining operations by using run-of-mine production and the revenue of the companies [6]. The study reveals the most significant keywords for the mining industry are to be ‘‘automation,” ‘‘robotics,” ‘‘Internet of Things (IoT),” ‘‘big data,” and ‘‘real-time data” based on the frequency analysis. However, when the relations between the different terms are considered (betweenness centrality and closeness centrality in network analysis), ‘‘automation,” ‘‘machine learning,” ‘‘Internet of Things,” ‘‘3D printing,” and ‘‘artificial intelligence” are the most relevant terms. In addition, the study also puts forward that, as the scale of operations gets larger the appraisal of digital technologies becomes more prevalent. Small-scale mining companies have less inclusion and adaption of smart tools, hardware, and software [7].

The fourth industrial revolution vision in the mining industry includes integrating smart systems, automation, and remote control into the operations that will undertake regular tasks or dangerous jobs in the work environment [2, 8]. The miner’s role is essentially to control the production with the real-time data flow from the machinery by engaging with the operators, suppliers, or experts remotely. Lööw et. al., [9] introduces the concept of Mining 4.0 as a notion corresponding to Industry 4.0 in the mining industry. In their paper, the authors outline future of mining through the perspectives of miners’ along with navigations for companies for a smooth transition. The envisioned miner within Mining 4.0, namely Miner 4.0, is described as an augmented miner with extended memory and senses supported by the technology. According to Lööw et. al., [9], Miner 4.0 is supported by technology to acquire information and improve situational awareness, which will aid them in extending their memory or sensory capabilities. In the study, it is stated that in this new working context, the required skills are more abstract and technical, where workers can learn or create values in the workplace. With this vision, lifetime learning and continuous education are effectively acknowledged to be a part of mining companies. Mining 4.0, smart systems, automation, and remote control will take over dangerous as well as routine work so that operators can focus on learning, creating, and valuing work tasks in a safe environment [9].

In the context of this evolving work environment, disruptive technologies have been imposing significant shifts on the job definitions, the number of jobs or the required skills, underlying the skills development, and the importance of transfer of skills to reduce the social costs for the community [10, 11]. It is imperative that all the stakeholders including industry professionals, business owners, and workers need to be prepared to optimize themselves with the slowly modifying automated industrial culture. Therefore, in order to identify and mitigate the disruptive effects of these impending changes in the mining industry, in this study, we conduct several interviews with industry professionals, who have been actively operating in the US mining sector, with the help of exploratory analysis and non-probability sampling [12, 13]. These methods are particularly useful in early stages of research, helping researchers identify patterns or themes within the scope of the research interest. As a limitation of the study, given constraints on the number and sampling of the participants due to time restrictions of the research, the employed convenience sampling method might introduce reduced generalizability of the findings. Based on the qualitative and quantitative analysis, we propose a novel job similarity metric that can be a baseline for human-centered, personalized, interactive job, and training recommendation system utilizing the publicly available O-NET database of the United States Department of Labor.

From this point forward, we present a literature study and our qualitative survey results in Section 2, followed by our methodology to calculate the gap between the current condition of the workers and the future work that they aspire to or the industry converges. We then evaluate our proposed methodology for some of the jobs our industry experts commonly identified to address how individual assessments can lead to a more effective job transition with the proposed human-in-the-loop process.

2 Literature

Through the third industrial revolution, over the past 30 years, the IT revolution has brought about a radical transformation of the world, both in our daily and work lives, which is comparable to the mechanization and electrification of the first and second industrial revolutions [3]. The concept of how technological change shapes skill demands is indeed a macroeconomic issue. Rapid computerization, task demands have dramatically shifted, reshaping the labor force supply trends globally. In his study, with the introduction of electronics and IT to achieve further automation through third industrial revolution, Acemoglu [14] inquires how the increase in computer availability changes the tasks performed by the workers and, thus, the associated skill demands. More importantly, he investigates the skill-complementarity to new technologies by design or by nature and whether there is a bias introduced to technology change by the rapidly increasing number of skilled workers. Acemoglu [14] highlights the evolving skill-replacing technological changes in the eighteenth century, pointing out the technology-skill complementarity phenomenon. At this point, not only the magnitude but also the direction of technological change needs to be elaborated as well. It is suggested that the market size determines the direction of technical change for different inventions. As the number of skilled laborers increases, the market for technologies that complement skills will get larger. Thus, more inventions will come in this context, increasing skilled labor productivity and their income. When there are more skilled workers, the market for skill-complementary technologies is larger. Therefore, according to the study, new technologies are by their nature complementary to skills, so there has always been a skill-biased technical change to some extent, and the recent past incorporated a rapid expansion of new technologies with an acceleration in skill-bias [14].

With a similar goal, Autor [6] also proposes an economic model that implies measurable changes in the composition of job tasks between 1960 and 1998, with the introduction of third industrial revolution, to investigate “how computerization alters job skill demands.” Although there have been studies to investigate the impact of computerization on expected skills that reveal reduced input for routine manual and cognitive tasks as well as increased labor input for non-routine cognitive tasks, the author examines what it is that computers do—or what it is that people do with computers—that causes educated workers to be relatively more in demand. With a motivation of setting a link between technical change and skill demand in economic literature, Autor [6] analyzes the temporal job task requirements using the Dictionary of Occupational Titles (DOT) and samples of employed workers from the Census and Current Population Survey of the United States. In the economic model they develop, the author investigates how the decline in computer prices changes the task demands within industries and occupations. The shift in jobs is regarded as a positive impact on the labor force, increasing the quality of the workforce capable of producing new tools and managing and guiding them with their skill set. At this point, it is important to note that The Occupational Information Network (O-NET), the utilized database in this study, is developed to replace the Dictionary of Occupational Titles (DOT) (United States Department of Labor, 1991). In order to keep the database current, a continual data collection process is performed with a goal to identify and maintain the changing characteristics of workers as well as jobs [7].

The economic perspective attempts to define, quantify, and project the required skills to embark on the future industry, as upskilling workers are related to employability, productivity, and economic growth. Similarly, the paradigm shift introduced by German National Academy of Science and Engineering as Industrie 4.0 (Industry 4.0) [3] is transforming organizational ecosystems and expected tasks from the workforce, incorporating the state-of-the-art technology that also evolves the expectancy from the human capital. Unlike traditional organizations, the decentralized setting will prevail to underline the importance of interaction, communication, or coordination within the workforce in this new organizational setting. The interconnected, dynamic, and collaborative nature of Industry 4.0 facilitates its integration into business ecosystems which will impact varying sectors to a greater extend, inevitably [4]. Given this shift, social acceptance of this new technology is central to adapting to this cultural change, emphasizing the importance of sociotechnical approaches that embrace the employees as the central component in the working environment. Within this new organizational schema, Operator 4.0 and Education 4.0 are the two social concepts of Industry 4.0, the former referring to the skillful workers supported by systems to improve their physical and sensory capacities, while the latter defines an AI aided, personalized education process that promotes a lifelong, time/location independent training within the scope of tailored learning formats [4].

In regards with the introduced training environments, Vavenkov (2022) underlines the change in training schema through the modern digital modeling technologies. The expansion digitalization in mining activities expands practical training applications not only for future mining engineers, but also for working specialists. High-quality simulations play a crucial role for presenting the working environments identically, which is made possible with the developments of virtual and augmented reality (VR-AR) solutions [15]. It is important that the simulation of the mining environment be of a high quality almost indistinguishable from the actual environment. In this context, the development of process solutions based on virtual and augmented reality (VR/AR technologies) is most relevant. In the same context, Young and Rogers (2019) emphasizes how VR mines also enable training applications to improve worker performances as well as to avoid dangerous conditions, which offers an invaluable contribution concerning risk management and safety practices [16].

Digital twins of mining and processing plants facilitate a smooth control through the virtualization of systems, process operations, and real-time data visualization that would improve decision-making processes. Computer-based simulation systems (CSS) have been incorporated for personnel training and qualification validation, resolving the problems in training process engineers and operators [17]. In line with that information, Alvarez et al., emphasize that active introduction and development in digital design education as well as mixed training systems are cardinal for mining engineering professionals, as they not only improve the educational processes but also enable the formation of information economy, based on expanded information technologies that would improve operational efficiencies [18].

Industry 4.0 is shifting the organizational ecosystems and the expected duties of the workforce, essentially reshaping the entire prospect of human capital. The asset of human capital is a combination of skills, education, and expertise that form a base for well-being and productivity in the work environment [4]. According to Meyer et al., [19], competence is the combination of attributes, abilities, skills, knowledge, and experience from a person, which are necessary for performing life and job roles [19]. The skillsets and competencies required to transition from one job to another are critical in adapting to the industrial shifts and shaping the future workforce. While a systemic, analytical approach to identify the current and past trends in skillsets reveals the impacts on disruptive forces in industries, the transferable and flexible skills that could be projected to future industry needs are more critical in reshaping and upskilling the current workforce [20]. The significance of a skill or ability in a certain sector is as crucial as its relevance for speedier adoption to future employment. Although data mining methods capture the past and current demands of the industry through text mining or keyword analysis, identifying how important each skill within the boundaries of industries’ expectations is critical for career transitions. Therefore, quantifying skill values that facilitate the transitions across job domains is needed to redefine and reshape the future workforce. In his study, Vista [20] uses network analysis where each node presents jobs and the connection transition from one state to another. The skills mediating the transitions between two jobs are considered to locate between the nodes where all pairs in the network can also account for each skill’s contribution levels (i.e., weight). After implementing permutations across pairwise job transitions by taking the importance of values of each contributing skill, the centrality of each skill is then calculated to reveal whether it is a cross-cutting one between different occupational transitions [20]. This ratio level network-based metric, which identifies the most relevant skill for most efficient transitions between two occupations, is called betweenness centrality in network analysis (Freeman, 1977, 1978 as cited in Vista, [20]).

Given this ongoing transformation, in this study, we aim to identify the present condition of digital transformation and digital acceptance in the mining sector, particularly in the USA, based on interviews conducted with industry professionals. In the light of the information and inferences acquired from the interviews, after inspecting the O-NET database carefully, we provide descriptions for some of the frequently mentioned occupations within the scope of expected-transforming jobs. In line with that information, we finally develop a metric to determine the critical gaps to transition from one job to another, that essentially quantifies as to whether a transition between the job pairs is feasible or not. In the upcoming section (Sect. 3), we elaborate on the methodology followed for both qualitative (Section 3.1) and quantitative analysis (Section 3.2). Along with describing the details of the methods, we evaluate our findings as we present them by setting a relationship between our interview results and data analysis. To validate our inferences, we select four different expected job transitions identified through our interviews, calculate the expected skill gaps between them, and associate these calculated results with our qualitative assessments. In Section 4, we present our results along with a comprehensive discussion of our findings.

3 Methodology

The study is developed to pave the ground for an AI system that makes use of a comprehensive set of featurized input of employees of mining industry in the USA, such as their skills, education, strength, interests, and job desires.

The first module of the study consists of conducting interviews with industry experts and analyzing our qualitative and quantitative survey reports (Appendix 1). We present the results of the interviews by elaborating on the consensus or different remarks of our participants. After implementing text-mining analysis on our survey results, we generate five classes: data-science, safety and environmental issues, business/management and mineral economics, technical/operational services, and traditional engineering services. After text classification, we evaluate these categorical changes in short-term (3–5 years), mid-term (6–10 years), and long-term (10 years and 25 years out) time periods comparing them to the current situation.

The second module of the study focuses on developing a job similarity metric by utilizing O-NET database of the United States Department of Labor [7]—which is a comprehensive database of the US job market—for our data analysis and algorithm development. In the final step, we analyze our developed method to the envisioned job transition scenarios commonly raised by our interviewees in the mining workforce.

3.1 Qualitative Analysis

In our conducted survey, we mainly focus on identifying the critical needs of the mining workforce, the competencies or knowledge base that the current workforce need to stay viable in their current position or to stay competitive in their work line. We conducted nine interviews with 11 industry experts acting as managers, consultants, directors, or engineers with at least 10 years of experience in the US mining industry. We also examine anticipated career shifts in the US mining sector; provide training to upskill the current workforce and the experts’ projections in the near/mid-future and in the long term. In addition to text analysis and classification of the survey results, we also conduct a semantic analysis based on our determined categories (i.e., data-science, safety and environmental issues, business/management and mineral economics, technical/operational services, and traditional engineering services) and contextual assessments.

The study is planned as an exploratory study with the focus group currently working in the mining industry, with a goal to achieve clear and quality information about the current state of the mining sector with the ongoing technological transformations. In line with the goal of exploratory analysis, the primary goal is to have an insight and develop a more concrete understanding of the digital transformation and its effects on the workforce in mining sector [12]. Given the research restrictions, convenience sampling is conducted for data collection [13]. Convenience sampling is defined as a non-probability sampling method, where the participants are selected based on their availability and accessibility. Although more rigorous sampling methods are suggested for increased generalizability, the method presents itself a feasible, time and cost-effective tool that enables data acquisition when especially a limited access/availability of the potential participants, as in this study [12, 13].

The general prospect of the interviewees is that the realm of automation of the machines, along with data analytics, machine learning, and artificial intelligence will increasingly influence the mining industry, meaning probably less workforce covering large geographic areas in fleets. The consensus is that there will be more and more shifts from human-driven technologies to machine-driven technologies in the upcoming years. Noting the ongoing shifts in the mining industry and the nature of the existing job standards [9, 11], the expectation is that for every job that is displaced, there will be a corresponding need for a person to program, configure, oversee, replace, install, deploy, build, and design.

From the perspective of mining operations, the general opinion is that the standards for recovering commodities from wastes are predicted to be far more the norm than is today, meaning mining mine waste repositories and solid waste landfills will become an increasing contributor to our material and mineral demands. Furthermore, a more rigorous understanding of water, waste management, environmental, and social responsibility processes will be needed in the sector for consideration with the expected higher standards of tailings management facilities and critical waste management repositories, environmental compliance, and/or stakeholder engagement programs. In addition to the waste management and environmental responsibility standards, the need to be more effective in addressing water and energy conservation, greenhouse gas, and carbon footprint issues will be more profound.

One of the most crucial aspects put forward by the experts is the cultural aspects of applying automation in mining operations. Customers’ strong culture of operational excellence in their current state is a hardship in adopting automation. That is, the adaption rate of miners to autonomy is regarded as comparatively low in the mining sector, mainly because of the cultural aspects of the industry. A simple example is, given that the business case is quite suitable for truck autonomy [21, 22], the interviewers state that there is still a reluctance to change in the industry as to when and how they fully adopt it.

3.1.1 The Insights on Critical Needs and Gaps

According to our survey, two major critical needs in the current workforce in the mining sector are underlined as building knowledge and adaption of autonomous mining and a robust understanding of data sciences. The reliance of mining industry on third-party consultants and companies is tied to these lacking knowledge in the mining sector. The experts indicate that the mining industry is just starting to embrace small robotic changes or digital transformation in vehicle intervention systems, such as collision avoidance for haul trucks, telematics, and tele remote operations for underground mining. However, these are accounted as the very first steps in autonomy. Therefore, having people who have strong analytical skills, data scientists, and employees who understand how software needs to function is critical in the workforce. Regarding staying competitive in the industry, data science with an understanding of how data can be used to optimize your mine environment and an understanding of autonomous mining—robotics is the future.

Another critical need is mentioned as the lack of “strong” generalists in the industry, although having many area experts. In order to stay competitive, our interviewees predict that employees will need to be equipped with programming skills, in addition to being able to manage different contexts and comprehending mining processes. Engineers looking for a balanced view between technical and business knowledge, security, computing, advanced analytics, and robotics will be more in demand. Above all, how technology can be leveraged to improve all business, personal and operational objectives with a higher emphasis on soft skills are put forward during our interviews.

According to our survey, willingness to keep learning and thirst to grow skills are the two major traits to stay competitive in their current jobs. These two qualities are underlined as a key to stay viable as an employee in the future workforce as well as in the global market. This is described more as a soft skill that initiates one’s self-motivation to learn more and go the extra mile when encountering a problem without being prompted. From the entry-level to high-level jobs, self-initiators who pursue learning and knowledge are characterized as critical competency.

Furthermore, business finance and how it relates to the introduced technology is underlined for both manufacturer’s and customer’s perspective by the industry experts. Understanding what current technologies are and what they bring to the mine site need elaboration on both ends, which means having a strong business case for the technology in need as to why the application is viable for the mine site. Competency for both ends is indicated as critical in our survey results.

3.1.2 Job Replacement Projections

According to our survey results, in the next 5 years, manual functions will be operated autonomously, most probably almost every operation in surface mining operations and mine surveying. The surveying and scanning tools via drones, drilling exploration, pre-feasibility before the mine permitting, blasting, rock movement, and haulage operations are some examples mentioned during the interviews. The common view is that more and more shifts from human to machine-driven technologies will be present in 3–5 years. For instance, surveying tasks have been increasingly carried out by aerial vehicles or automated robots, as they are now effective tools for mapping and surveying tasks [23,24,25]. This ongoing change inevitably reshapes the traditional surveying jobs to digitally generating and analyzing data, creating transferrable digital plans required by these machines. These jobs are described as engineering-type jobs, as in site engineering, mining engineering, and civil engineering realms across our interviews. The machines will mostly work in the direction of those types of employees, meaning a higher job demands for these reshaped realms in the upcoming years.

The displacement of many truck operators, including drilling jobs, is expected in the upcoming 3–5 years. The number of equipment operators (production and haulage), process plant operators, and most surveyors (largely unskilled and semi-skilled jobs) are expected to be reduced and replaced by computer-operated systems, requiring skilled computer technicians and engineers for each job title. Accordingly, the growth of people that understand how to be a controller or dispatcher within a mine site and these skills will be in very high demand. Mechatronics engineers or technician job definitions are predicted to shift how to troubleshoot and repair a technological piece of equipment, such as a problem with the brake system versus a problem with the control system that is telling the brakes to actuate. Specifically, on the operator and technical side, a lot of current strong skills are envisioned to be in need as to how the systems work and interact with each other, such as hydraulic systems, electrical systems or the GPS, satellite, and cellular monitoring systems. Since one individual is likely to be able to operate more than one piece of equipment remotely, the operator jobs are expected to be fewer, requiring new immediate skills to accomplish the same jobs. Some jobs will inevitably stay on-site and somewhat the same: mechanics, electricians, and welders, for example, but the tools of their trade will inevitably become more automated, requiring the workforce to be more technologically savvy in addition to their trouble-shooting and technical capabilities. Additionally, on the job creation side, application engineers (or specialists), focusing on efficient operational application of technology and really extracting the value, which is equipped with more skills in data analytics or the cultural side of the application, is expected. Additional roles such as communications technicians and engineers will be in high demand as a networking system is required to connect all these machines together and keep the networking system alive in the future.

3.1.3 Training Applications and Transformations

The projections about the training converge around a dramatic digital platform shift in training, from instructor-led to training on demand, to digital delivery distance education or virtual-augmented realms. In the mining industry, current training operations are based on instructor-led, classroom-based training, some computer-based training, and on-site training. Although there is simulator work being done on mechanical simulators for operators to some extent, the future of training is predicted to expand a lot more VR-AR haptic; in a controlled environment, many pieces of machinery operating or performing a particular task in the process plant, without physically touching it.

The inteviewees predict that the augmented reality realm is going to allow us to tie subject matter experts that are not on the job site or not near the machine, whether it is an operator or technician. Having witnessed the change in the work ecosystem some companies already have been emphasizing system evaluation, understanding how the machines work, communicate, breakdown or any of the systems that remotely monitor, help diagnose and repair. These companies conduct on-the-job (OJT) technical solution expert training and application engineering training to upskill their current workers based on these expectations.

According to our survey results, some companies immediately develop training for familiarization (for executive level and business development preparation as well), and they offer some in-depth training to increase the skillsets of those proficient on the previous standards. However, to stay in demand on both employee and employer’s side, there is a need to conduct more training customized to industry needs instead of generic technological training. To improve processes, upskilling the employees is needed to leverage technology in every step, incorporating customized training concepts for specific problems. One common remark constantly underlined during interviews is “As jobs are displaced, other jobs will be created.” This general view emphasizes the importance of conducting training programs in the industry for testers of technology, remote operators, drone operators, app designers, and maintenance workers based on the expected competencies.

About the training delivery, currently, some mines use digital twins, so you have the real operating function of that mine that provides a training environment [15, 17, 26,27,28]. In relation to future training applications, the common idea is that people will not attend in-person learning at all while they learn on their own device, at their own pace in the near future, underlining the fact that they already utilize VR simulators for some of their training on operator side. However, an important concern raised during our interviews is that some companies utilize online training systems to check some boxes so that the company is not at fault from a legal perspective. Obviously training programs that were delivered with such motivation are ineffective as they are conducted for the sake of compliance. As important as it is in technical services, physically being on a mine site or classroom to have an experience on operating machines, safe blasting operations or collaborative learning is neither a requirement nor a long-term medium for training purposes.

3.1.4 Future Projections

According to our interviews, conversion of surface mines at the end of their life to underground mines and greenfield mines starting in underground and deeper mines will be prevalent as the reserves run out in the upcoming 15–25 years. Along with that, environmental concerns and human condition issues will be much more prevalent in the mining industry. As resources/ore bodies are exhausted, terrestrial mining will transition more to offshore deposits and off-terrestrial deposits, which is a shift to even realistic space mining.

In the mine environment, a huge amount of electrification in all original equipment manufacturer (OEM) products is anticipated, which means a shift from diesel to hybrid systems. The workforce will be much more scattered, and companies will be able to do more work with a workforce that is either through particularly companies’ available workforce or through on-demand type of force. Furthermore, advances in water treatment technology, as well as advances in renewable energy technologies, are predicted to make energy more economical in mining settings (i.e., concentrated solar, organics-to-energy, engineered, and run-of-river hydropower).

Industry 5.0, as a new paradigm of industrialization, is summarized around certain goals and visions such as human-centric collaboration with AI and robotics, sustainability, net-zero operations/manufacture, and circular economy [29, 30]. In comparison to Industry 4.0, this newly emerging paradigm or vision of Industry 5.0 considers human/workforce dimension not only as a better supply of optimized, connected, and utilized resource as in the classical human resources approach; it actually envisions a human-centric working ecosystem around future workers. Here, the transitions mentioned by the experts will also be aligned with this vision and also care about the well-being of the worker when we consider the features of emerging or expected technological shift from in-site, sometimes manual and risks—exposed aspects of labor to remotely controlled machines and automation. Another area for potential improvement in the industry could be in the handling of explosives. While the expert participants of the study provided no specific insights, collaborative robotics and remote-controlled operation notions of Industry 5.0 could be utilized to reduce risk exposure, which is also, aligned with human centric and worker well-being oriented principles of this paradigm. The technological changes envisioned by industry experts also consistent with the features of Industry 5.0, such as the “circular economy” or net-zero goal-based operations. The transition to electrification emphasized by the experts is another underlined aspect that aligns with this vision or dimension of Industry 5.0.

As corroborated by the expert inputs, the transition and emerging complexities are not only a generational issue but also a cultural one. As they suggest in mining, there might be even more organizational resistance compared to the manufacturing industry. For example, the delayed adoption of driverless trucks is regarded as cultural issue, which was pointed out by one of the interviewees. This observation suggests the need for relevant leadership and organizational change efforts to facilitate a smoother transition to future stages of industry evolution.

3.1.5 Preliminary Analysis of O-Net Database Based on Interviews

O-NET Database of the United States Department of Labor is a comprehensive database of the US job market that includes a rich set of variables describing work and worker characteristics, including skill requirements in the form of a wide range of work and job-oriented data categories [7]. The developed content model introduces a framework of work classification taxonomy in the USA with a defined set of occupations and the related data for each one. The model includes both occupation-specific (using job descriptors) and worker-specific (focusing on worker-oriented descriptors) categories, which is transitionary in nature, enabling data utilization across jobs, sectors, or industries through cross-occupational and occupational-specific descriptors. These attributes fall into several categories either focusing on the key features/characteristics of workers or occupations, namely worker characteristics/requirements, experience requirements, occupational requirements, workforce characteristics, and occupation-specific information for data analysis and algorithm development [7].

With our goal to understand expected job shifts in the mining industry, we aim to reveal transforming occupation patterns within the industry through the insights of our participants and highlight immediate training needs for the mining workforce. Figure 1 shows the frequency analysis of our stacked interviews after tokenization and removing stop words. The figure implicitly puts forward critical issues in the scope of our questions, enabling to quickly visualize these significant remarks made by the experts (i.e., training, machine, customer, subject matter expert, etc.).

Fig. 1
figure 1

Frequency analysis of stacked interview texts

Full size image

Although several occupations were mentioned during our interviews, we present the commonly referred occupational shifts to observe their corresponding skills in detail. Firstly, for the technician group, truck operators, surveyors, and machine operators are frequently mentioned on the replacement side, while technicians who operate similarly to surveying engineers who collect and store digital data; telecommunications techs or radio mechanics who install data communication lines or fixing them are the common technician occupations on the creation side. For the engineering group, sales engineers who are technically and technologically able to promote OEM products and logistic engineers who supervise the operations in the new mine environment are the emphasized two engineering pathways on the job creation end. To probe into these occupations, we inquire about the skill needs for potential career shifts to demonstrate the common skills with their importance and levels for each of them as shown in Table 1. 

Table 1 Top ten important skills for some mining jobs
Full size table

 Preliminary data analysis on commonly mentioned transitioning jobs were conducted with O-NET Database data, simply filtering the top ten important skills for the desired job and simply calculating the absolute skill differences between the current and desired jobs with L1-norm. L1 normalization normalizes each element by dividing by the sum of the absolute value of all elements. L1 normalization is as follows:

$${F(p)}_{i}={~}^{{p}_{i}}\!\left/ \!{~}_{{\sum }_{j}^{{N}_{t}}{|p}_{j}|}\right.$$

where F denotes the L1 normalization and p = \(\left[{p}_{1}, {p}_{2},...,{p}_{{N}_{t}}\right]\), and \({p}_{{N}_{t}}\) is the maximum value [31].

Table 1 shows top ten skills required for some mining jobs, which might be a transitioning path for mining-industry employees in line with the foreseen changes in the job types. As seen from Table 1, from merely a skill-based, one-dimensional perspective, there are certain commonalities between some potential career paths, while there are significant skills that need to be acquired for a potential transition. In the O-NET database, the importance scores ranges between [0 and 5] while the skill score levels ranges between [0 and 6.2] calculated based on the occupational expert ratings. For the detailed calculation of skill scores, readers are advised to review the documentation of the O-NET database webpage [7] [https://www.onetcenter.org/dl_files/AOSkills_ProcUpdate.pdf].

For a better understanding, in Fig. 2, we plot the top ten skill of importance and their corresponding levels for some of the commonly mentioned jobs by our experts. As an example work transition scenario, for industrial truck and tractor operators, there are six common skills with surveying technicians, with a total skill level gap of 16.12 (as mentioned earlier, calculated using Manhattan distance between the job pair); while there are only four common skills with remote sensing technicians, with a total gap score of 27.11 (Fig. 2). However, surveying technicians and remote sensing technicians have eight common skills, showing an easier potential transition between the latter pair, with a total gap score of 11.87. Likewise, in order to transition from a mining engineer position to a sales engineer position, although there are six common skills between the two, there seems to be a skill gap score of 15.01. On the other hand, the number of common skills is seven with the logistics managers as far as the top ten required skills are concerned, with a calculated skill gap of 9.27 (Table 1 and Fig. 2). At any rate, these skill-based comparisons provide a sense of effort for a desired career transition that might simply reshape one’s expectations throughout the potential change.

Fig. 2
figure 2

Skill importance and skill levels for some mining industry jobs. a Industrial truck and tractor operators. b Surveying technicians. c Remote sensing technicians. d Mining and geological engineers. e Sales engineers. f Logistics managers

Full size image

Given the projections on the expected replacement of the current workforce, it is apparent that these predetermined skill transition paths could present the potential career change paths for the current workforce. While a systemic, analytical approach to identify current and past trends in skillset reveals the impacts on disruptive forces in industries, the transferable and flexible skills that could be projected to future industry needs are more important in reshaping and upskilling the current workforce [20].

For further elaboration, we utilize text analytics to filter and preprocess our survey data. We tokenize and classify the interviews to capture a better understanding of major response categories. We create five main classes: data-science, safety and environmental issues, business/management and mineral economics, technical/operational services, and traditional engineering services, to categorize the data and analyze the trend in these defined categories for different time-frames provided in our questionnaire ((i.e., current, short-term (3–5 years), mid-term (6–10 years), and long-term (10 years and 25 years out) projections) (Fig. 3). In the first part of our survey, we attempt to elaborate on the current critical need of the workforce, the primary competencies needed to stay viable in the mining industry, and the trainings currently provided to educate the workforce. The second and third sets of questions examine the expected competencies/knowledge base from the workforce to stay competitive in the industry, anticipated shifts in the occupations in the near/mid future, and the training needed to upskill the workforce in line with the expected shifts. Finally, in the last group of questions, we ask our interviewees about their future projections 10 and 25 years out if it is different from their already made forecasts within the previous categories.

Fig. 3
figure 3

Text classification results across different terms. a Current. b Near-term (3–5 years). c Mid-term (6–10 years). d Long-term (15 years and out)

Full size image

Figure 4 shows overall categorization of survey results. According to the classification results, 35% of the mentions in our interviews fall in data science and autonomy class, followed by 26% in technical and operational services. From the contextual standpoint, operation-related remarks are stated in the scope of expected job shifts in the industry. Following that, business and mineral economics mentions come forward with a score of 20% and 12% for the safety and environmental issues. The general view is that there would be an increase in understanding the impact of the operations both safety-wise and environment-wise in the upcoming years. Increase in personal and environmental safety awareness will be central from exploration to operations (Fig. 4).

Fig. 4
figure 4

All categories through all terms

Full size image

Our temporal analysis shows that (Fig. 3) the critical competency need in data science and autonomy category increases across all terms, starting from 28% in current demands, reaching to 43%. Across all terms, there’s a close percentage for the mentions of technical/operational services, with a decrease in the engineering-related mentions in far-future, which are probably related to the expected job shifts in these categories, with a close relation to autonomy for both services (i.e., some traditional engineering tasks operationalized by technicians).There is a consensus among the interviewees that in the long run, autonomous change will be operationalized effectively.

In the same way, the business and mineral economics category shows similar scores for the different terms, which is often brought forward in relation to upper and mid-management, but also desired from entry job positions to capture their role in the whole operation as well. This is because the mining operations are extensive in scale, and the operation efficacy is critically important for minimizing the production losses for each level; therefore, monetary loss from the business standpoint. In addition to these outcomes, there seems to be an increasing trend in the safety and environmental issues (from 7 to 16%), in line with the expected business model transition in mining operations throughout the world. Socially responsible mining operations have been inevitably more and more profound across the world, where community engagement is significantly essential, and participatory engagement of the communities is present. The common view is that operations will be more sustainable and greener in the upcoming future, which companies should consider and plan in order to be able to operate worldwide (Fig. 3).

Based on this ongoing transformation, there is an apparent need in policies that should account for skills development in the mining industry for transferring technology skills and knowledge, by both companies and governments. These policies should specifically be established and encouraged by government incentives; because when these incentives are driven by industries, not by the government, the potential of the suppliers for a strong supply chain is not assured. According to Ramdoo (2018), the initiatives focus on improving the suppliers’ capacity, specifically on mining industry-related technical challenges, their long-term plan might not be aligned with the national strategies (Ramdoo, 2018). Therefore, future worker training is not only a local issue concerning the individual industries but also a strategic economic administration in terms of adaptation of current workers to this evolving work environment and job security of young generations, which is closely related to social license to operate as well (SLO).

3.2 Quantitative Analysis—The Proposed Metric

The conceptual foundation of O-NET database is structured upon a content model that offers a framework that entails the most important types of work, and combines them in line with theoretical and empirical organizational literature, divided into two major groups: worker-oriented and job-oriented categories. Through these categories, the framework provides a holistic embodiment of work profiles. The database entails descriptors for categorical job information organized through Standard Occupational Classification System (SOC), with specific elements defining each. O-NET database contains over 1000 occupations with detailed SOC requirements supplying a rich set of information for developers and companies, describing work and worker characteristics. The database also contains occupation-specific information to specify training, develop position descriptions, or redesign jobs that would be utilized to reshape training paths [7].

Knowledge is a body of information that an employee should have at the time of the hiring; the skills are the proficiencies gained through training or education (basic skills, social skills, complex problem-solving or resource management skills, technical skills). In strong relation to the two, ability is the capacity to apply existing knowledge and skills simultaneously to actualize a given task or activity (cognitive, physical, psychomotor, and sensory abilities) [7].

Our preliminary analysis reveals that although there have been some company-based strategies, there is a critical gap in determining the future workforce training in the US mining industry to identify shifting needs of existing jobs or for the new jobs that will be created for a smooth job transition. In determining the mutual compatibility of a current and desired job, in this part of the study, we propose a metric that combines worker requirements; characteristics accounting for skills, knowledge, and abilities needed; and work activities describing required for each occupations. After the qualitative analysis of our interviews, we used our proposed metric to validate the individual cases that were pointed out in our interviews. Our approach introduces three main goals, which are setting a prototype for a human-centered job recommendation system, prioritizing the needed skills for personalized training programs. In our view, the proposed metric will help in facilitating not only the recruitment processes but also retraining paths for employees, which will simplify the HR recruitment operation to a great extent, therefore saving significant resources.

3.2.1 Proposed Similarity-Metric

The Occupational Information Network (O-NET) program is a primary occupational database comprising a measurable, standardized set of descriptors that define occupations in the USA. As mentioned earlier, these descriptors are organized in a framework that reflects worker-specific and occupation-specific characteristics. Based on our preliminary studies, discrete analysis of each individual feature of occupation provides rather limited information, emphasizing the need of a combinatory and continuous metric to define each job. Therefore, in this study, in order to quantify dissimilarity/similarity between occupations, we propose a metric that takes knowledge, skills, abilities, and work activities into account in an equally weighted fashion.

We utilize Kullback–Leibler Divergence (KL Divergence) [32] algorithm to analyze the relative entropy between the occupations of interest. The algorithm measures how one probability distribution is different from another probability distribution, which can be described by the following equation:

$$\text{KL}(P||Q)= \sum P\left(x\right)\text{log}(\frac{Q(x)}{P(x)})$$

where, \(P\) and \(Q\) are probability distributions of a certain domain of the current occupation and the desired occupation. In addition, \(P\) and \(Q\) are defined on the same probability space, \(x\), as elements of a certain domain [32]. While deciding the parameters for the metric, we take the content model reference of O-NET database into account. Worker characteristics (i.e., abilities), worker requirements (i.e., skills), knowledge, and occupational requirements (i.e., work activities) are major constituents of the utilized content model reference of O-NET database. As a model-case implementation, by accounting for abilities, knowledge, skills, and work activities equally as significant parameters defining the occupation of interests (it might be customized based on the unique worker/employer profiles), we define a gap parameter between the current occupation and desired occupation as follows:

$$\text{Gap}=\frac{1}{4}({\text{KL}}_{\text{Abilities}}+{\text{KL}}_{\text{Knowledge}}+{\text{KL}}_{\text{Skills}}+{\text{KL}}_{\text{WorkActivities}})$$

where, \({\text{KL}}_{\text{Abilities}}\), \({\text{KL}}_{\text{Knowledge}}\), \({\text{KL}}_{\text{Skills}}\), and \({\text{KL}}_{\text{WorkActivities}}\) are the KL divergence of abilities, KL divergence of knowledge, KL divergence of skills, and KL divergence of work activities. The Gaps are normalized, and the similarity metric between the occupations is calculated as follows:

$${\text{Gap}}_{\text{Normalized}}=\frac{\text{Gap}-{\text{Gap}}_{\text{min}}}{{\text{Gap}}_{\text{max}}-{\text{Gap}}_{\text{min}}};\text{ Similarity}=1- {\text{Gap}}_{\text{Normalized}},\text{ Similarity}\in [0 1]$$

The following five thresholds were then applied to the similarity scores for grouping transition hardship and better visualization.

Similarity [s]

Value

\(0\le s\le 0.2\)

0.1 (very hard)

\(0.2<s\le 0.4\)

0.3 (hard)

\(0.4<s\le 0. 6\)

0.5 (medium)

\(0.6<s\le 0.8\)

0.7 (easy)

\(0.8<s<1.0\)

0.9 (very easy)

\(s=1\)

None value

Table 2 lays out the first ten abilities, skill, and knowledge for the job title of “Mining and Geological Engineers, Including Mining Safety Engineers” based on O-NET database. These categories are also provided with their corresponding importance level for each specific job title, which is essentially required for recruitment.

Table 2 Ability, skill, and knowledge for the job “Mining and Geological Engineers, Including Mining Safety Engineers”
Full size table

3.2.2 Applications with the Proposed Metric

After defining our metric, we first utilize general group of occupations to separate discernable similarities and discrepancies to test our approach. As seen in Fig. 5, in-group transitions are expectedly easier, while the between-group shifts are categorically harder. However, finely tuned group thresholds enable tracking the difference within the group variations as well. The generated heat map delineates between group differences, pointing out how difficult it is to change career paths across jobs, for instance whether it is easier for a logistic analyst to become a sales manager or financial manager based on proposed metric (Fig. 5).

Fig. 5
figure 5

An example of gap analysis between different job groups

Full size image

Once the algorithm is tuned with different job families, we implement mining-related occupations to identify the similarities within the selected jobs and set up potential transition paths for current workers in the mining industry. Fig. 6 illustrates the heat map for selected mining-related jobs (i.e., engineering, technician, and management jobs). To make the heat map more recognizable, we include sociologists, geographers, and anthropologists in our job group and generate the map for dissimilarity scores; the darker the color scale, the lower the similarity. Based on the visual inspections, an initial assessment of each job family and individual occupational transitions can be inferred from our generated heat map. For example, the map highlights the hardship in transitioning from technician jobs to management positions with a clear boundary or isolates some engineering jobs from other as a dissimilar option for some management jobs. For instance, for an electronics engineer, “Computer and Information Systems Management” or “Industrial Production Management” is an easier transition than “Marketing Management” based on our scores. Likewise, as far as the technician job family is concerned, “Industrial Production Manager” and “Operations Manager” positions are better career options than other management jobs (Fig. 6).

Fig. 6
figure 6

Similarity map for mining-related occupations

Full size image

Figure 7 shows the engineering and management similarity matrix of the occupation of interests, along with their corresponding scores for each transition, including sociologists, anthropologists, and geographers for better contrast. In this focused map, it is seen that for a mining engineer, logistics or marketing management is an easier transition rather than industrial production management, with similarity scores of 0.83, 0.82, and 0.68, respectively. Similarly, given the scores, civil and environmental engineers can transition to a “Chief Sustainability Office” position comparatively easier than any other engineering job title in the group. “Mining and Geological Engineers,” on the other hand, have a better chance to move to a “General and Operations Manager position than “Mechanical Engineers or Electrical Engineers” based on the similarity indexes. Likewise, for an environmental engineer transferring to a chief sustainability officer position instead of an operational manager or operation manager job is an easier option based on our defined similarity score. Therefore, the calculated scores can guide workers who desire to change their career paths or have to go through a career change in line with the ongoing digital transformation.

Fig. 7
figure 7

Similarity map with scores for engineering and management jobs

Full size image

Furthermore, with our selected test set consisting of 40 occupations, we plot the KL divergence metrics for these referred jobs to observe their distribution behavior across other jobs. Figure 8 shows the distribution of KL metrics for occupations frequently emphasized in the interviews, either on the creation or displacement end (i.e., truck operator, remote sensing technicians, surveying technicians, etc.). The mean, standard deviation, and kurtosis of these distributions reveal whether and to what extent that job is transitionary in our test set. Based on these values, among these six plots, logistic managers and mining/geological engineers have a better stand in changing their career paths (lower mean, lower standard deviation, and kurtosis) than other occupations in Fig. 8.

Fig. 8
figure 8

KL divergence histograms for some occupations of interest in mining industry

Full size image

In sum, our proposed metric that makes use of several occupation-specific characteristics presents itself as a functional tool, with its inclusion of several job-specific properties and its basis of probability distribution, rather than relying on discretized-values. As the career changes of an employee are reshaped during their professional life, in-group and between-group job transitions are each critical. Through this proposed tool, employees who lack a measured ability, knowledge, and skillsets or interest to change his/her job will be able to get an objective list of how s/he should proceed on a personal scale in that path. In addition, the distribution range can also be used to indicate whether a transition is possible or comparatively easier from that particular job to others.

In addition, from the employer’s perspective, this system will facilitate to understand the characteristics of their current workforce and to plan for better workforce training programs. Given that, the O-NET database is a readily available and reliable source, our proposed tool presents itself as a basis for a personalized AI system that would be deployed with individual data collected.

4 Results and Discussions

The ongoing energy transition, meeting environmental, social, and governance (ESG) goals, and decarbonization advances innovative approaches in mining with increasing digital and autonomous operations. This would lead to small-scale companies that cannot compete with large-scale industry partners laying off their workers due to competition. The capacity to transform to digital will effectively determine which company will stand along with who would be employed by these companies and stay competitive.

The industrial revolution creates skill shortages or mismatches in the labor market with a consequential negative impact on the economy, leading to a demand for a creative and strategic workforce that could function in cross-functional activities or processes [11]. In his study, Lööw [9] defines the miner’s role as essentially controlling the production with the real-time data flow from the machinery by engaging with the operators, suppliers, or experts remotely. The technology supports the workers in acquiring information and improving situational awareness, which will aid them in extending their memory or sensory capabilities [9]. According to the study, there is an apparent deficit in the diffusion of digital technologies in smaller operations. Larger mining operations appear to select and apply digital technologies suitable to their needs, whereas operations with lower production rates do not implement the currently available digital tools and hardware technologies to the same extent. These minor producers may require other digital transformation solutions tailored to their capabilities and needs and applicable to their scale of operations [9].

In this study, with the goal of generating a personalized and customized data analysis hub to mitigate the disruptive effect of automation, we propose a novel metric to quantify how a job is different from another based on required ability, skills, knowledge, and required job activities in the mining industry. We implement qualitative and quantitative data analysis to get a more comprehensive understanding of our interviewees. Through the proposed novel metric, we evaluate the forthcoming job transitions by using the O-NET database. In developing this tool and inferential analytics, how difficult or feasible it is to transition from one job to another is evaluated for the mining industry’s future workforce.

Our preliminary data analysis on the interviews reveals that the critical competency need in data science and autonomy presents an increasing demand across all terms, from 28% in current demands to 43%. While technical/operational services preserve similar percentages, engineering-related mentions show a decrease in the far future. These trends in both categories are interpreted as directly related to the expected job shifts in these categories, closely related to autonomy for both services. The consensus of our participants is that in the long run, autonomous change will be operationalized commonly and effectively.

The analysis reveals that the in-group job transitions are significantly easier than the between-group transitions, which proves our initial assumptions. For instance, if a worker currently acting as an “Excavating and Loading Machine Operator” desires to move within the operator and technician category to a surveying technician job, s/he has comparatively a less challenging transition than an “Extraction Worker” to do the exact career change. Our metric provides similarity scores in relation with each potential transition within the worker’s expectancy space. The similarity metric also reveals which career plan is more feasible among other options, proposing individualized career paths for job seekers.

Through our proposed metric, we also generate the KL-metric distributions for each job of interest. The distribution itself is also representative of each job profile as it pertains significant information to each desired job profile as well. A systemic, analytical approach to quantifying current and past skillsets trends will also provide the opportunity to identify transferable and flexible skills, which are critical in job transitions and upskilling the current workforce as well. The study proves the applicability of the extensive O-NET database for identifying job transitions in the US mining sector, benefiting from expert interviews and qualitative and quantitative analysis. Depending on the level of interest and adaption by the industry, the O-NET database stands as a great resource in the prediction of personal career paths as well as what’s needed for training for future workers in the mining industry.

The job similarity index enables a methodical examination of job responsibilities, detecting significant similarities and variances based on skills, knowledge, and competencies, therefore enhancing workforce adaptability. This tool facilitates the transfer of workers to new positions that are becoming more crucial in a digitally evolving environment hence aiding the industry’s move towards more sustainable and technologically advanced practices. The index can facilitate targeted training programs by assessing the disparity in skills between present employment roles and those that are projected to be in high demand. This guarantees that the labor force is equipped not only to adjust to novel technology but also to participate in energy-saving and reduced-emission activities, which are essential in the energy transition. The index aids in developing focused training programs that meet specific requirements identified using job similarity criteria, hence facilitating upskilling and reskilling. Training in automation, digital skills, and other competencies pertaining to energy-efficient mining technology might be included.

The job similarity index provides valuable insights that can assist policymakers and industry leaders in formulating plans to guarantee that the mining workforce is adequately prepared to fulfill the requirements of the energy transition. These plans include regulations pertaining to workforce growth, incentives for training, and recommendations for implementing sustainable mining operations. The index facilitates job mobility by providing workers in the mining industry with valuable information about future career transfers, enhancing their professional growth. Ensuring mobility is crucial for keeping talent and enabling the workforce to adjust effectively to the industry’s changing demands.

Although this study has limitations on generalizability for qualitative assessments, it presents itself as a valuable application for how qualitative and quantitative approaches could be useful for future worker training in the mining industry. It also presents the value of publicly available O-NET data for planning job transitions in the sector. Keeping limitations in mind, our findings report that the technological change in the US mining industry is reshaping the operational expectancies in the sector more and more, calling for a dynamic re-training system in the companies and systematic governance within the sector. Looking for a solution as to how the workforce transitions should be managed, we offer the O-NET database as an invaluable resource, customizing and setting out the changing job expectancies within the mining companies, which can be incentivized by government agencies as well.

5 Future Work

By collecting industry-specific-individualized data, an AI tool can be developed and continue to learn as end users engage with the system, creating a data hub for the future workforce in the mining industry and employers. O-NET database and custom questionnaires are essential to training this AI network, improving the prediction rates based on each employee profile through personalized data collection in the mining industry, considering not only technical skills but also personal interests as well.

In the deployment phase, the AI system is planned to generate individualized career paths for job seekers and offer training needs for companies. Such a system can act as a home base for the mining, metals, and manufacturing industry in the US job market.