Potential Scenarios and Hazards in the Work of the Future: A Systematic Review of the Peer-Reviewed and Gray Literatures

Abstract It would be useful for researchers, practitioners, and decision-makers to anticipate the hazards that workers will face in the future. The focus of this study is a systematic review of published information to identify and characterize scenarios and hazards in the future of work. Eleven bibliographic databases were systematically searched for papers and reports published from 1999 to 2019 that described future of work scenarios or identified future work-related hazards. To compile a comprehensive collection of views of the future, supplemental and ad hoc searches were also performed. After screening all search records against a set of predetermined criteria, the review yielded 36 references (17 peer-reviewed, 4 gray, and 15 supplemental) containing scenarios. In these, the future of work was described along multiple conceptual axes (e.g. labor market changes, societal values, and manual versus cognitive work). Technology was identified as the primary driver of the future of work in most scenarios, and there were divergent views in the literature as to whether technology will create more or fewer jobs than it displaces. Workforce demographics, globalization, climate change, economic conditions, and urbanization were also mentioned as influential factors. Other important themes included human enhancement, social isolation, loneliness, worker monitoring, advanced manufacturing, hazardous exposures, sustainability, biotechnology, and synthetic biology. Pandemics have not been widely considered in the future of work literature, but the recent COVID-19 pandemic illustrates that was short-sighted. Pandemics may accelerate future of work trends and merit critical consideration in scenario development. Many scenarios described ‘new’ or ‘exacerbated’ psychosocial hazards of work, whereas comparatively fewer discussed physical, chemical, or biological hazards. Various preventive recommendations were identified. In particular, reducing stress associated with precarious work and its requirements of continual skill preparation and training was acknowledged as critical for protecting and promoting the health and well-being of the future workforce. In conclusion, the future of work will be comprised of diverse complex scenarios and a mosaic of old and new hazards. These findings may serve as the basis for considering how to shape the future of work.


Introduction
The nature of work, the workplace, and the workforce are changing rapidly, differently, and to a greater extent than in years past, affecting greater numbers of individuals in profoundly powerful ways (Daheim and Wintermann, 2016;Johansson et al., 2017;Jain et al., 2018;International Labour Organization (ILO) 2018;World Bank Group, 2019). Waiting until the effects are fully revealed to address them could be highly problematic and inadequate. Instead, anticipatory thinking regarding new and persistent factors affecting work and workers is required. Rantanen (1999) was one of the earliest investigators to detail future of work scenarios relevant to occupational safety and health (OSH). He foresaw a focal shift from traditional health outcomes, such as mortality and hospitalization, to multifactor determinants of health (e.g. behavioral, environmental, social) and indicators of functional capacity (e.g. work ability, work motivation, and quality of work life). Rantanen (1999) suggested that future scenarios would involve not only current and ongoing hazards but also new ones, yet to be observed.
The current review provides a comprehensive characterization of the future of work in terms of the scenarios that may occur, the hazards that may result from them, and recommendations to address them. It also offers solutions from the literature to address the hazards associated with the future of work.

Methods
A three-pronged approach was used to conduct a comprehensive search for literature describing future of work scenarios and the hazards associated with them.
Prong 1: searching for future of work scenarios in the peer-reviewed literature Using the principles of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA; Moher et al., 2009), the authors conducted a search for peer-reviewed literature containing future of work scenarios. Searches were conducted between April and December 2019 using ABI-Inform, Business Source Complete, EconLit, Embase, Ovid, Public Health Database, PubMed, PsychINFO, SCOPUS, Social Sciences Database, and Sociological Abstracts databases. Initial inclusion criteria required a reference be published in English between 1999 and 2019 and contain in either its title or abstract at least one keyword from the lists generated by the authors (see Supplementary Appendix I). After removing duplicates, the search yielded 1202 records for screening. The authors' process for screening and reviewing these records is detailed in Supplementary Appendix II. The most critical screening criteria included ensuring an article satisfied the current study's operational definition of work scenario: a product (i.e. piece of writing) that either 'describes some possible future state or tells the story of how such a state might come about' (Bishop et al., 2007, p. 8). Screening also included an automated search for key terms, a manual review of abstracts for mention of at least one future of work scenario, a refinement of the included publication years (2009 to present), and the level at which work was described (job or task). The authors then worked in pairs to manually complete a critical review and assessment of each article that passed the screening process. Figure 1 provides an overview of the selection process, which resulted in the retention of 17 peer-reviewed articles.
Prong 2: searching for future of work scenarios in the gray literature The search for future of work scenarios in the gray literature entailed government websites, nongovernment websites (business, industry and academic), Google Scholar, and Google. Initial inclusion criteria required a reference be published in English between 1999 and 2019 and contain in either its title or abstract/text at least one keyword from the lists generated by the authors (see Supplementary Appendix I). If needed, Google searches were further refined using search options for file or site type. All gray literature cited in a recent future of work review issued by the ILO (Balliester and Elsheikhi, 2018) was also added to the initial web search results. In total, 269 records were retrieved. These records were manually screened and reduced to a set that (i) was published in 2009 or later (to mirror the revised peer-reviewed inclusion criteria); (ii) made projections or predictions about the future of work; and (iii) contained at least one term from the following list: scenario, model, hazard, risk, and expos* (e.g. exposure, expose, exposed, etc.). Four documents contained scenarios that clearly predicted or speculated risks or hazards and were retained for analysis.
Prongs l and 2 yielded a surprisingly small number of articles and reports, suggesting the need for additional sources to fully explore future of work scenarios. To accomplish this, an additional search method (Prong 3) was designed to broaden the search for scenarios and hazards.

Prong 3: Supplemental Searching for Scenarios, Hazards, and Themes in the Published Literature
The authors conducted supplemental literature searches of the peer-reviewed, gray, and general literatures to further expand information gathered in the systematic searches. The searches were completed using backward and forward snowball approaches (Wohlin, 2014) as well as ad hoc searches of relevant literature to December 2019. Seven recent future of work literature reviews, found in the initial scan of literature, served as the 'start set' for this search: Ballister and Elsheikhi (2018), McKinsey Global Institute (2017), Acemoglu and Restrepo (2018), Peruffo et al., (2017), British Safety Council (2018, ILO (2019), and Schulte and Howard (2019). Papers were included if they pertained to future of work, included at least one keyword from the search list in Supplementary Appendix I, and were consistent with the identification of a future scenario  (rather than commentary) as described by Spaniol and Rowland (2019). The remainder of this article synthesizes future of work scenarios; summarizes anticipated hazards associated with the future of work from peer-reviewed, gray, and supplemental literatures; discusses major themes considered relevant to the future of work by the authors (see Supplementary Appendix III); and provides recommendations for anticipating and responding to challenges associated with the future of work.

Future of work scenarios and hazards in the peer-reviewed literature
This review identified 17 peer-reviewed papers describing future of work scenarios, summarized in Table 1A. Generally, these are conceptual macro-level scenarios describing broad characteristics of the politicoeconomic context or attendant labor market. For the purposes of review, these papers are combined into four non-mutually exclusive scenario categories describing a convergence of concepts: (i) changing patterns of employment and work organization, (ii) management of technological change and human-robot interaction, (iii) OSH challenges, and (iv) ethical issues. Figure 2 quantitatively summarizes four major hazard categories based on scenarios found in peer-reviewed sources. The most common scenarios involved the exacerbation or creation of psychosocial hazards, followed by physical hazards. Chemical and biological hazards were rarely represented in the scenarios.

Changing patterns of employment and work organization
Based on the five papers in this category, technology, globalization, demographics, and urbanization were identified as contributing to changing patterns of employment and work organization (Johansson et al., 2017;Bellace, 2018;Caruso, 2018;Ghislieri et al., 2018;Traulsen and Druedahl, 2018). Overall, the scenarios entail workplaces where smart machines, materials, warehouses, and other factory systems will continuously exchange information with human workers, resulting in a reduced need for human workers. These automated facilities are projected to cause a qualitative knowledge transformation-from bodily and tacit into more theoretical and abstract knowledge and skills and from craftsman-like qualifications to more technical qualifications (Johansson et al., 2017;Caruso, 2018). Caruso (2018) offered an evidence-based critique on the promises of knowledge work and the positive effects of digitalization, noting that work organization has become more precarious rather than more horizontally integrated, and workers have yet to experience increased decision-making power or autonomy. Additionally, work has become more creative for only a fraction of highly skilled workers, and the distinction between work-time and life-time has weakened. Bellace (2018) further noted that the needs of workers are becoming delinked from employment, and a new paradigm for labor law is needed. This new paradigm should emphasize the changing nature of work. Moreover, Bellace concluded that in the 21st century, the notion of employment that undergirds labor law has been breaking down. As we move into the information age, labor unions have been weakened by a decline of class consciousness and technological changes.
In this category, future work will be an evolution of the current era. This is depicted as an industrial revolution (Industry 4.0) and is seen to differ from previous ones because it involves exponentially paced technological changes with the potential to cause systemic disruption in most, if not all, industries (Ghislieri et al., 2018). The revolution has been described in publications from various countries as an inchoate transformation of production of goods and services resulting from applications of a new wave of innovations (Caruso, 2018). Such digital innovations and the emergence of Industry 4.0 may constitute epochal social changes. One such change may involve labor relationships and the individualization of the relationship between workers and machines, which will affect unions' bargaining power and workers' collective actions (Caruso, 2018). The context of work in the future may be further influenced by the eroding of the 20th-century notion that employers have obligations to employees. The advent and growth of platform work illustrates this change, with companies such as Uber and Lyft declaring they do not 'employ' their workers (Bellace, 2018).
Hazards to workers depicted in these scenarios are related to technological developments and include changing employment patterns, precarious work, unemployment, underemployment, competency obsolescence, psychosocial stress, and work intensification. The increased incorporation of technology (e.g. robots, machines, digitalization) maintains the potential to increase the complexity of production systems across industries, which may increase workload and stress levels and decrease worker motivation (Johansson et al., 2017;Latos et al., 2018). The risk of addiction to new workrelated technologies and pressure to constantly be online and available 24/7, in an 'always on' working world, may also adversely impact the well-being of workers and their families (Ghislieri et al., 2018). Additionally, Unclear OSH ethics must be developed in response to the changing world of work.

Chemical-Nanomaterials.
Psychosocial-Blurred work-home boundaries; control and monitoring by employer; disconnect between organizational performance economically (turnover, costs, profits) and socially (ethical responsibilities); discrimination; globalization, market intensification; robotics in the workplace; social isolation. Consider privacy, ethical, social, and other implications when formulating policy.
Foster collaboration between OSH and other professionals. Integrate personal, professional, and institutional ethics. Introduce ethics courses to medical curricula. Mine existing data sources. Review the competence of OSH professionals. Healy et al.
3 Unclear Technology will impact jobs and skills in Australia, resulting in optimistic, essimistic, and mixed scenarios for the future. in terms of well-being rather than accident rates. Regulate and oversee work practices. Review regulatory frameworks and assess them for their adequacy in controlling changing risks in a changing workplace. Use fiscal incentives rather than sanctions to encourage behavior change, especially for small-and-medium enterprises. the growth of digitalization may impact gender patterns of workforce participation. The relationship between gender and new technology worker qualifications and identity could also change as technology is introduced into historically male-dominated workplaces and industries (e.g. mining) as physical strength is replaced by technological enhancements that will enable more women workers (Johansson et al., 2017).

Management of technological change and human-robot interaction
According to the seven papers in this category, robots were identified as participating in many of the future work environments (Qureshi and Syed, 2014;Murashov et al., 2016;Koppenborg et al., 2017;Islam, 2018;Pham et al., 2018;Pyke, 2018;Djebrouni and Wolbring, 2019). Key drivers for increased robot use include increasing labor costs and worker shortages (Qureshi and Syed, 2014). Both developed and developing countries are amenable to digitalization and automation, which fuels automation anxiety and causes workers to fear their contributions will become obsolete because of new technologies (Islam, 2018;Pham et al., 2018;Pyke, 2018). Some of the scenarios in this category (e.g. Pham et al., 2018) speculate many workers will lose their jobs to workplace technologies, such as robots. Certain groups, such as people with disabilities, may benefit from new technologies while also being placed at increased risk of occupational marginalization because of robotics and human enhancements (Djebrouni and Wolbring, 2019). Although some non-scenario studies (e.g. Frey and Osborne, 2013;Baert and Ledent, 2015;Peruffo et al., 2017) have predicted as many as 30-47% of jobs are at risk of automation, automation at the task level rather than occupation level is likely more realistic (Autor et al., 2003).
The interaction of humans and machines-particularly robots-will depend on whether and how humans accept robots. Questions of trust may limit human acceptance rates for robots. Human factors are an important consideration in the design of robots as collaborative human-robot work environments become more common (Koppenborg et al., 2017). For example, Johansson et al. (2017) observed a dearth of evidence quantifying the degree to which humans perceive, accept, and treat robots as replacements to human coworkers. Murashov et al. (2016) pointed out the labor force's longstanding experience with robots dating to the 1970s, which has influenced the development of various national and international standards for the safe design and use of robots, on which future policymakers can lean. The main hazards related to robots in the future of work are projected to be mechanical, electrical, thermal, and noise, with traumatic injuries being the greatest physical health effect expected (Murashov et al., 2016).

Occupational safety and health challenges
This category-comprised of four papers-provided commentary on OSH challenges associated with anticipated future of work scenarios (Hauke et al., 2018;Leso et al., 2018;Chia et al., 2019;Niu et al., 2019). Health and safety recommendations for issues highlighted below can be found in the Recommendations in the Literature section of this article. Leso et al.'s (2018) review of 22 papers describing health and safety issues related to Industry 4.0 emphasized adverse psychological outcomes as the primary health concern for workers. Key issues and concerns in the technology-driven workplace of the future include mental overload, work density, a mismatch of workers' skills and task requirements, privacy invasion, reduced human-to-human contact, and unemployment (Leso et al., 2018). New types of accidents may result from a lack of guidance or standards for interfacing with new technologies, and workers may experience increasing tension between the virtual and the real world with increasing technological integration in the workplace (Leso et al., 2018). Chia et al. (2019) presented a useful overview of the need for a new workplace safety and health strategy for Industry 4.0. They identified new hazards to well-being, including job displacement and blurring work/home boundaries. Their conclusion suggests current workplace safety and health strategies in developed countries, premised on a quantitative risk management model and traditional full-time employment, will be inadequate considering technological advancements and changing employment relations. They also noted the complexity of the human-machine interface may leave workers unable to cope with the intricacies of technological products and result in failure to use all their available functions. Other noteworthy hazards associated with Industry 4.0 will include psychosocial stressors from 24/7 digital work platforms, physical health risks related to nanoparticles and hazardous volatile organic compounds from additive manufacturing, and biological hazards from virulent and pathologic organisms in synthetic biology. These rapidly developing hazards may increase concern over the adequacy of surveillance systems for adverse effects in the future. Hauke et al. (2018) conducted an online survey of 398 labor inspectors in Germany about future hazards for a variety of industries and occupations. Work intensification topped the list of hazards generated by these subject matter experts. They also expressed a belief that extension of responsibility (due to staff shortages) will go together with increased work intensity in the future. which provide the opportunity to develop a new wave of OSH management to address 'stagnant management' worldwide.
Ethical issues Iavicoli et al. (2018) focused on ethical concerns in the changing world of work. They identified key drivers and barriers that will influence ethics for future OSH research and practice. Factors such as globalization, demographic changes, and increased technology use and advancements may bring new ethical challenges to the OSH field. In this future of work scenario, it may become increasingly important, yet difficult, to balance the wants, needs, and interests of workers (e.g. healthful and safe work and the right to be informed about risks) with those of organizations (e.g. maximizing production, rights to industrial secrecy) and the common good (e.g. community-level health, safety, and well-being).

Future of work scenarios and hazards in the gray literature
Future scenarios in the gray literature are summarized in Table 1B and were generally described in terms of business models, labor market changes, social values, and category of work (i.e. manual or cognitive). Main topics of these scenarios included technological impact on work, worker control, national economic outlooks, and worker capabilities. The World Economic Forum (WEF, 2018a) identified eight future of work scenarios based on the rate of technological change and its impact on business models (steady or accelerated), the evolution of learning among the current and the future workforce (slow or fast), and the magnitude of talent mobility across geographies (low or high). The future of work will be defined by combinations of these variables, with different scenarios playing out simultaneously in different areas, industries, age cohorts, and socioeconomic groups (WEF, 2018a). OSH implications of these scenarios will vary based on the level of uncertainty generated and the resultant mental and physical health effects for workers.
The European Foundation for the Improvement of Living and Working Conditions (Eurofound, 2018) identified three broad scenarios for the future of work that combined applications of digital technologies to economic processes. The three categories included the automation of tasks, digitalization of processes, and coordination of platforms, which are considered social vectors of change (Eurofound, 2018). Automation of tasks will eliminate some types of work and create others. Critically, the work must be automated at the task level rather than the job or occupation level. This is particularly true for jobs comprised of non-routine and intellectual tasks, which represent a large portion of work today and would lead to high rates of displacement and unemployment if fully automated (Eurofound, 2018). Digitalization involves change in the work environment and the nature of work processes. Digitalization of economic processes raises some serious concerns for workers' autonomy, privacy, employment stability, income levels, work schedules, and work-life fit. Coordination of platforms is a direct response to changing employment conditions and regulations. A multiplatform system allows for the division of labor into very small tasks that can be tedious and repetitive, which is not an ideal psychosocial work environment and can be associated with feelings of alienation (Eurofound, 2018). Brown et al. (2018) surveyed 10 029 people and identified four mega trends to explicate the future context for work. The trends represented poles on two axes: collectivism ('fairness and equality dominate') versus individualism ('me first rules') and business fragmentation ('small is powerful') versus corporate integration ('big business rules all'). Additionally, to better compete in these scenarios, 70% of respondents indicated that they would use 'treatments' to enhance their brain and body by 2030 if it would improve their employment prospects. Degryse (2016) evaluated the rapidly evolving scenario of a digital economy and identified various hazards such as lack of job security, disruption of work-family balance, work intensification, stress, burnout, the virtualization of relationships, and the confusion between what is urgent and what is important. The issue of increased training of workers was identified as central to success in the digital economy.

Future of work scenarios and hazards in the supplemental literature
A seminal paper by Williams (2008) framed the issue of devising scenarios by identifying a common narrative structure that underpins a multitude of contrasting visions of employment in the future. These narratives adopt a similar storyline, in which all employment is classified into one of two categories (one of which is more favorable than the other) and then ordered into a temporal or normative sequence. In contrast to others' views, Williams argued for a multifaceted understanding of the future that recognizes heterogenous and multiple directions of employment. Commonly, many other scenarios used the 2 × 2 approach (Schultz, 2007;Ponce del Castillo and Meinert, 2016;Hajkowicz et al., 2016;Korge, 2018;Dellot et al., 2019).
Useful supplemental literature was identified and characterized (Table 1C). Using focus groups and a Delphi method, DeBruyne and Gerritse (2018) identified various scenarios of the future of work and their implications for the future physical workplace. Their findings suggest most workers will spend more time outside a traditional office environment by 2025, splitting their production time between the office, home, and elsewhere. This is due, in part, to the notion that future work will be characterized by a reduction in the amount of work that must be accomplished in a company. Instead, work will be comprised of more complex processes and human tasks that must be digitalized, namely knowledge-intensive and service-oriented work. This work will require collaboration with other employees or stakeholders in the supply chain.
Hazards identified by DeBruyne and Gerritse (2018) include the danger of being connected 24/7 and the deterioration of social cohesion within the organization. In addition, labor contracts are expected to shorten, and the number of guaranteed and long-term jobs is expected to diminish. The employment relationship in the future will no longer be connected to a specific function, but rather to where work makes a meaningful contribution to one or more projects, teams, or employers. Increased flexibility will have favorable or unfavorable consequences for work-life balance, and favorable and unfavorable impact on workers' experiences of burnout and their physical and emotional health (Bell and Blanchflower, 2018;DeBruyne and Gerritse 2018).
In a report from the European Agency for Safety and Health at Work (EU-OSHA), Stacey et al. (2018) developed four scenarios of the future of work, focusing on digitalization and new emerging risks associated with them. The overall conclusion was that digitalization and related technologies will have a profound effect on the world of work resulting in more complex and diversified work environments (see Table 2). The four scenarios fit into cells of a 2 × 2 matrix where one axis is 'governance and public attitude' (low to high) and the other 'economic growth and technology application' (low/resistive to high/ supportive). The four scenarios are labeled: Evolutionhalf of jobs have changed, with 10% replaced by automation; Transformation-most jobs have changed, with 50% replaced or fundamentally changed by automation; Exploitation-half of jobs are fully automated, with high unemployment; and Fragmentation-20% of mainly lower skilled jobs are fully automated.
More generally, the Reserve Bank of Australia identified four categories of work based on underlying skill context: (i) non-routine cognitive, (ii) non-routine manual, (iii) routine cognitive, and (iv) routine manual to characterize jobs (Acemoglu and Autor, 2010;Healy et al., 2017) see Supplementary Appendix IV. Analyses of Australian data showed a shift away from routine employment towards non-routine employment over the last 30 years, demonstrating that the distribution of work is changing. Considering the empirically supported 'routinization hypothesis', which posits that technology is displacing workers performing routine tasks, Healy et al. (2017) compared four reports from Europe (Méda, 2016), the UK (UKCES, 2014), the USA (Shift, 2017), and Australia (CSIRO, 2016) to imagine and contrast multiple future scenarios, simultaneously. Although these reports differ vastly in their purposes, they can be categorized as having optimistic, pessimistic, or mixed (majority) scenarios. Optimistic scenarios described a world of full employment and environmental sustainability, whereas pessimistic scenarios depicted a world with fundamentally weak labor markets, fewer worker protections, and large social inequality. The mixed scenarios revealed a world in which technological change has driven the automation of work, including skilled jobs, to the point of widespread technological unemployment and fragmented tasks, resulting in societal and political pressure to reduce the negative effects (Healy et al., 2017).
In another analysis, Daheim and Wintermann (2016) described the findings of an international Delphi Study by the Millennium Project, in which three highly specific alternative visions of the future were presented. All three scenarios assumed rapid technological change and radical impact on work, where in-demand skills will be unlike those touted today and human-machine cooperation will be more widespread. The three scenarios ranged from depicting accelerated technology, to social and economic change, to political and economic turmoil, to self-actualized economies throughout the world. Subramony et al. (2018) described important hazards in a future scenario for service workers by 2050. The scenario was predicated on a transformation of traditional employee ongoing organizational relationships and work arrangements to a clearly less defined employee relationship, where an employer promotes opened-ended inducements to control the process by which work is performed. In line with this observation, other investigators noted that 'the constant search for the next gig might be a source of anxiety as it involves a sense of job security that engenders emotional examination' (p. 966), (Cappelli and Keller, 2013). This scenario increasingly relies on transient just-in-time work, which raises multiple concerns regarding the nature of work and the effect on worker well-being, including dehumanization of work and increased stress due to job insecurity and periods of involuntarily unemployment. The use of AI in Europe may be viewed as a hazard scenario that was analyzed for the ethical issues involved in promoting trustworthy AI (HLEG, 2019). The report of the scenario identifies the principle of preventing harm as critical to consider when deploying AI. Moreover, it concluded that the OSH implications for workers, especially vulnerable groups, must be transparently understood as well as the overall technical robustness and safety and the impact on human autonomy. This report is useful because already there is nascent evidence of hazards of unfair treatment and discrimination (Moore, 2019).
The human-machine interaction scenario for AI-enabled technologies was described by Howard (2019), who noted that 'systems controls which are not fully understandable to humans or fully responsive in practice as they were in design can lead to negative consequences.
[…] Some accident analyses may be biased to safeguard the integrity of the technical system at the expense of the nearest human operators' (Howard, 2019, p. 921;Elish, 2019). Also, Janssen et al., (2019) identified 'mode confusion' as a critical factor that can influence workers' trust in automation and operation of it. Mode confusion occurs when the abstraction of information is insufficient for the user to anticipate a machine's state (Maeda and Ushio, 2017). In the coming years, human interactions with automation are expected to be the subject of mode confusion (Janssen et al., 2019).
Just as past industrial revolutions initiated a broad pattern of industrialization that led to sweeping social and political change, so too is AI emergence likely to significantly influence the global economy. 'The integration of AI technologies across human society could also spark a process of cognization analogous to changes wrought by industrialization' (Scharre et al., 2018, p. 3). Hazards to workers are not inherent in AI but may arise from how it is implemented (Moore, 2019).
An ad hoc and snowball search of the scientific and gray literature led to the identification of critical themes that will influence the future of work. The themes pertain to the following factors: technological, demographic (young, older, women, migrant, and disabled workers), temporal (time, work, and leisure), global, urban, climate-related, human enhancements (cognitive, physical, worker monitoring), hazardous exposures, advanced manufacturing, biotechnology synthetic biology, sustainability, and political and economic factors. These themes are summarized in Table 3 and described further in Supplementary Appendix III.

Inequality as an underlying factor in future scenarios
The concept of unequal distribution of wealth, income, opportunity, gender, race, and access to information underlays many of the scenarios in the gray and supplemental literature (e.g. Daheim and Wintermann, 2016;Shift, 2017;Stacey et al., 2018;Brown et al., 2018). Clearly, there is a broad awareness and growing discourse on the negative consequences of inequality in the world (Piketty, 2014;Joyce and Xu, 2019). There is a growing consensus that in the future, the changing nature of work may exacerbate these inequalities. Inequality, at the individual or population level, has important consequences for health, which can affect employability, job maintenance, and job opportunities (NASEM, 2017).

OSH implications of the future of work
The various scenarios and projections of the future of work synthesized from the scientific, gray, and supplemental literature identify many hazards with OSH implications as summarized in Tables 1-3. The future of work and its impact on workers will be a mosaic of longstanding hazards that currently exist (e.g. exposures to chemicals, physical, radiological and biological agents); slips, trips, and falls; musculoskeletal hazards; repetitive work and manual material handling; longstanding hazards in new jobs (e.g. psychosocial stress due to technological displacement); and new hazards in new jobs (e.g. collisions with robots, discriminatory monitoring of workers through wearable sensors, and human-machine role ambiguity; Murashov et al., 2016;Peckham et al., 2017;Badri et al., 2018;Leso et al., 2018;Pot, 2018;Stacey et al., 2018). Figure 3 provides a strategy for assessing these old and new hazards in existing and future jobs.
The unflagging pressure of technology on workers, as manifested by the increasing pace and intensity of work and the expectation for humans to be able to function effectively with robots and in response to algorithms, could have devastating consequences (Marchant et al., 2014;Degryse, 2016). Flexible labor markets and the decline of trade unions may put future workers at risk of decreased job protection, whereas automation, lack of skills, and the inability to fill job vacancies may put as many as 400 million workers worldwide at risk of job displacement (MGI, 2017). Furthermore, workers displaced by technology will not necessarily be the same workers hired for new jobs created by technology. All these changes are predicted to lead to the development Table 3. Critical themes in the future of work literature a

Technological inevitability
Technology alone will not shape the future of work; social context dialogue and process are key (Winner, 1998;Perez, 2002;Little, 2008;Lemieux, 2014;Autor, 2015;Curry, 2015;Susskind and Susskind, 2015;ILO, 2017a;Creticos, 2018). Young workers Many children that entered school in recent years are projected to work with skills that do not yet exist. The demand for advanced cognitive and behavioral skills will increase and the demand for narrow job-specific skills will continue to decrease. (Krueger and Kuman, 2004;Council for Work and Health, 2014;Ederer et al., 2015;Cunningham and Villasenos, 2016;McGuinness et al., 2017;World Bank Group, 2019). Older workers Workers aged 55 and over (during the next few decades) will become one of the fastest growing segments of the workforce in many countries (ILO, 2018). Countries with higher rates of projected aging generally have larger proportions of older workers at risk of automation (Paton, 2014;Foresight, 2016;Basu et al., 2018;Harris et al., 2018;Healy and Williams, 2018;McGowen and Corrado, 2019).

Women workers
Women and men may experience technological job displacement differently since women traditionally perform more routine cognitive tasks. Women need to have more access to technology and training (Brussevich et al., 2018;World Bank Group, 2019).

Migrant workers
In the future, there is likely to be more than 160 million migrant workers globally. Migrant workers frequently experience increased rates of morbidity, mortality, and injury. (ILO, 2017b;IOM, 2017;Flynn and Wickramage, 2017;ESPAS, 2018).

Workers with disabilities
Future of work literature is linked to research focused on those living and working with disabilities. New technologies may both help and discriminate against persons with disabilities (Department of Work and Pensions, 2017;Kanady, 2018).

Social isolation and loneliness
Decentralized work may lead to social isolation and loneliness. Loneliness may have an impact on mortality equivalent to smoking 15 cigarettes per day (Rook, 1984;Murthy, 2017;Holt-Lunstad et al., 2010;Jeffrey et al., 2017;Nemecek, 2018;Ozcelik and Barsade, 2018;McMillan, 2019;Patel et al., 2019). Globalization A new phase of globalization is likely but will still be a determinant of job loss (due to 'off-shoring'), which is associated with adverse health effects (Benach et al., 2004;Kawachi, 2008;Dobbs et al., 2015). Urbanization In addition to worker safety and security indirect factors such as wage polarization, housing costs, and need for social services will be critical issues for how the future of work will be impacted by and impact urbanization (OECD, 2014;DuPuis et al., 2016;WEF, 2018a). Climate-related factors Outdoor work will be hotter. Working capacity of heat-exposed workers is expected to decrease, while deaths and illness in workers exposed to heat is expected to increase (Dunne et al., 2013;Kjellstrom et al., 2014;Schulte et al., 2016;Clayton et al., 2017;Maitre et al., 2018;Rigaud et al., 2018;Sylla et al., 2018;Cho, 2019;Dong et al., 2019). Cognitive enhancement Increased use of cognitive enhancing drugs is likely. Drugs that enhance cognitive capabilities (and that may also have physical effects) may be inappropriately promoted instead of work organization improvements (Dale and Bloomfield, 2016;MarketWatch, 2018). Physical enhancement Growing use of physical enhancements (such as exoskeletons) may increase the incidence of deleterious effects (Academy of Medical Sciences, 2012;Federici et al., 2015;Young and Ferris, 2017;Zingman et al., 2017;McGowan, 2018;Butler and Gillette, 2019;Hargreaves et al., 2019).

Worker monitoring
Wearable sensors or other monitoring technologies may result in violation of privacy and autonomy and lead to discrimination (Bandodkar et al., 2016;Moore, 2018;EU-OSHA, 2019;Zuboff, 2019). Advanced manufacturing Advances that involve changes in the process of manufacturing may present new hazards and old hazards in new settings (Hassall, 2015;Geraci et al., 2018;Pomeroy-Carter et al., 2018;Roth et al., 2019;Wang et al., 2020). of an array of psychosocial factors that can, in turn, lead to physical and psychological health effects (Benach et al., 2004;Koranyi et al., 2019;Rönnblad et al., 2019). These hazards will arise from work tasks, the precarity of work, and job insecurities, and are relevant not only for workers sufficiently employed but also for workers classified as either underemployed or unemployed (Janlert, 1997;Dooley, 2003;Dorling, 2009;Bjorklund et al., 2015;Benach et al., 2016).
Indeed, for the first time in the modern era, economic growth is becoming detached from employment. Brynjolfsson and McAfee (2014) refer to this as the 'great uncoupling'. It is unknown whether the unemployment that will result from technological and economic change will be a transitional or long-term effect, though some populations will become unemployed. The cohort of displaced workers and their dependents may need social protection and mitigating action. Long periods of unemployment may lead to depression, anxiety, poor self-esteem, divorce, substance abuse, increased chronic diseases, suicide, and mortality (Henkel, 2011;Marchant et al., 2014). What is more, unemployment may result in a lack of meaning in the lives of workers even if income substitution options can be implemented. The elimination of work itself may also be a source of dehumanization, despite the economic premise that favors leisure over work (Mokyr et al., 2015).
Increasingly, investigators are aware of the need to assess the hazards workers face over their life course to account for the dynamic relationship between work, nonwork, and health that evolves over time (Janlert, 1997;Burdorf, 2012;Amick et al., 2016;Degryse, 2016;Serra et al., 2017;Schulte et al., 2017).
To that end, the concept of 'worker well-being' is becoming the current and future goal of the OSH field (Anttonen and Rasänen, 2008;Allen, 2014;Schulte et al., 2015;Chari et al., 2018). Therefore, it is critical to assess the hazards and precariousness due to jobs and between jobs in one's working life, and essential to consider the integration of work and nonwork (Standing, 2011;Schulte et al., 2015Schulte et al., , 2017Amick et al., 2016;Bell and Blanchflower, 2018). This will be especially important in the future of work, given evidence to date on the impact of changing jobs, being displaced from jobs, and not being able to get a job (Degryse, 2016;Rönnblad et al., 2019). Indeed, along with the well-established physical, ergonomic, chemical, radiological, biological, and other hazards, many of the determinants of worker well-being are and will be expected to be psychosocial in nature and related to the quality and availability of work (Degryse, 2016;DeBruyne and Gerritse, 2018;Jain et al., 2018;Leso et al., 2018;Stacey et al., 2018).

Recommendations in the literature
The future of work will be a mosaic of scenarios of old, current, and new jobs and hazards (see Fig. 3). Below, the authors outline select key recommendations found during the course of the review process (see Tables 1  and 2). The looming question is what can be done to shape the future, today. Overall, there will need to be collaborative efforts among stakeholders and other decision-makers on the implementation of measures to ensure a smooth and safe transition to the future (ILO, 2017a; Badri et al., 2018;OECD, 2019). Critical in this quest is how to prevent and manage psychosocial

Hazardous exposures and disease
There are a vast number of chemicals in commerce and millions of workers with exposures to them (Calvert et al., 2012). Exposures may put them more at risk of occupational cancer may lead to occupational cancers in the future. There maybe a shift of exposures to developing nations. Also, physical, biological, radiological, and musculoskeletal factors also can be hazardous to large number of workers (EU-OSHA, 2007;Watanabe et al., 2011;Haagsma et al., 2012;Valencia, 2013;UNECE, 2015;Madhav et al., 2017;Fritschi, 2019;Wang, 2020;Siemiatycki and Rushton, 2020;Shearer et al., 2020). Biotechnology and synthetic biology Biological processes may become a major source of economic growth. Creation of new or altered life raises significant concern about potential health effects and ethical issues (Hewett et al., 2016;NAS, 2017;Howard et al., 2017;Gomez-Tatay and Hernandez-Andreu, 2019). Sustainability Many future workers may seek jobs that practice or address sustainability. Focusing on sustainability may be a new approach for advancing worker safety and health (OSHA, 2016). However, green jobs may have hazards (Brundtland, 1987;Bradbrook et al., 2013).

Political and economic factors
How societies organize and conduct themselves will favorably or unfavorably influence working conditions and the future of work (Walters and Wadsworth, 2014;Lippel et al., 2017;Chandy, 2016;Pyke, 2018;Kinder, 2019  risks, which are already superseding physical health risks and are of growing concern for the future workforce (Degryse, 2016;ILO, 2017a;Stacey et al., 2018). Badri et al. (2018) also identified 12 recommendations to foster this transition successfully. Among them were efforts to conduct research on psychosocial risks, prevention through design, and research on emergent risks at all levels of production. Another common recommendation found in the literature to help remedy key potential hazards workers will face is to provide them with lifelong and future-ready adult learning, training, reskilling, and upskilling (Bradbrook et al., 2013;British Safety Council, 2018;Stockton et al., 2018;OECD, 2019;Work Bank Group, 2019). Also, universal basic income has been raised as an intervention for precarious or nonstandard work (Hoynes and Rothstein, 2019).
Various investigators identified preventive measures related to future hazards. Howard (2019) concluded that 'a proactive approach to AI and its implications for the future of work requires OSH professionals develop strategic foresight to anticipate and prepare for the possibilities and challenges of AI-enabled technologies on worker safety, health, and well-being' (p. 922). Similarly, Leso et al. (2018) identified the need for risk assessment at the process design phase. They also promoted the need for international standards and adequate-specific training of workers. The EU-OSHA published a model of new and emergent risks to evaluate work settings. To complement and clarify the EU-OSHA (2013) definition, Fernandez and Perez (2015) developed a more refined risk model applicable to advanced manufacturing processes.
An RSA survey of British Members of Parliament revealed concern for the impact of technology, but 43% of respondents indicated lack of knowledge regarding how to influence it (Dellot et al., 2019). The RSA report on the future of work explains that we do have choices: 'We can choose to establish a robust regulatory regime for technology and data rights […] to create a tax system that shifts the burden onto those with the broadest shoulders […] to overhaul our education system so that we treat lifelong learning more seriously…[or] to create a competition policy that stands up to the power of large firms when they impinge on the wellbeing of workers' (Dellot et al., 2019, p. 3  Another approach by EU-OSHA (Stacey et al., 2018) identified various strategies that could mitigate the OSH challenges of digitalization and ICI-enabled technologies: • The development of an ethical framework for digitalization and codes of conduct • A strong 'prevention through design' approach that integrates a user/worker-centered design approach • Collaboration between academics, industry, social partners, and governments on research and innovation in developments by information and communication technology-emerging technologies (ICT-ETs)/ digital technologies to properly take account of the human aspects • The involvement of workers in the implementation of any digitalization strategies • Advanced workplace risk assessments, using the unprecedented opportunities offered by ICT-ETs, while also considering the full range of their possible impacts in terms of OSH challenges, as identified in this foresight project • A regulatory framework to clarify OSH liabilities and responsibilities in relation to new systems and new ways of working • An adapted education system and training for workers • The provision of effective OSH services to digital workers A critical next step is identifying mechanisms to accurately predict which jobs and tasks will change and disappear, including the timeframes in which these changes will occur (Chang and Huynh, 2016;Peruffo et al., 2017). In 2013, Vasic and Billard (2013) urgently called for a definition of robots, which may replace workers, and for specific safety guidelines to be addressed by the scientific and industrial community. They identified autonomous vehicles and mobile robots as two of the most urgent areas where safety guidelines are needed. The International SOS Foundation (2018) assessed what significant OSH changes need to be prepared for by 2030. Based on a global survey, key results revealed that 84% of respondents believed there will be an increase in complexity of health and safety requirements. Four major global factors include the (i) UN 2030 Agenda for Sustainable Development and its Sustainable Development Goals; (ii) revised Occupational Health and Safety Global Reporting Initiative (GRI) Standard; (iii) the ISO 45001; and (iv) ILO Guidelines on Occupational Safety and Health Management Systems. More than half of the survey respondents anticipated that health, safety, and environment will be a board level or 'C-suite' role by 2030. New scenarios and hazards and a growing incidence and prevalence of noncommunicable disease will present future challenges to the OSH field. To meet these challenges, the field may require a holistic 'biopsychosocial' approach (one that addresses the interconnection between biology, psychology, and socioeconomic factors) to promote health, well-being, and workability (Harrison and Dawson, 2016). Peckham et al. (2017) suggested that the changing nature of work requires a reconceptualization of occupational health in the future and a more holistic and public health-oriented model addressing worker health. The broader concept of worker well-being also emerged as one that will be considered in the 21st century (Chen and Cooper, 2014;Schulte et al., 2015;Litchfield et al., 2016;Peckham et al., 2017;Hudson et al., 2019).
The ILO (2019) called for a more 'human-centered' agenda for the future of work and safety and health to address not only hazards in a single job but also along the work-life continuum. To this end, in January 2019, the ILO Global Commission on the future of work called for a Universal Labour Guarantee, including fundamental workers' rights, an 'adequate living wage', limits on hours of work, ensuring safe and healthy workplaces, as well as the recognition of safety and health at work as a fundamental principle and right at work.

Limitations
The current study was designed to characterize future of work scenarios and hazards described in the published literature rather than offer a critical review of the extant literature. As such, it was delimited to sources from the peer-reviewed and gray literature that included descriptions of the future, which met Bishop et al.'s (2007) definition of 'scenario'. Critical studies and commentaries focused on present OSH conditions or trend extrapolation were not included. However, to ensure the current study's literature review strategy captured a comprehensive set of future of work perspectives, author affiliations were assessed for the 36 peer-reviewed, gray, and supplemental scenario sources. The affiliations included a considerable mix of academic/professional disciplines (e.g. OSH, economics, business management, engineering, and social science) and countries of origin (e.g. USA, Western Europe, Australia, China, Saudi Arabia, and Singapore), offering no evidence of any obvious skew or bias in the alternative futures that were reviewed. There were, however, notable gaps in the future of work scenarios that were identified by the three-pronged search methodology. Generally, the scenarios were described in broader labor market terms. They frequently provided nonspecific depictions of work in the future, describing tasks in terms dichotomized axes (e.g. 'manual or cognitive'; 'routine or non-routine'). The scenarios also included a lack of attention to topics frequently considered critical in future projections, such as climate change, infectious disease, social isolation, and the burden of chronic disease. Rare, too, was the inclusion of worker voice or input in the visioning and development of the future of work. Recent events illustrate the significant lack of attention to pandemics in the future of work literature. Although the future of work is driven by powerful forces such as technology, demographics, and globalization, the current COVID-19 pandemic could have a modifying effect by accelerating trends already underway (McGowan 2020). These trends include isolation of many workers, exacerbating inequalities among workers, and displacing workers from jobs. Though there is little empirical evidence on the effect of pandemics on work and the economy, there have been many predictions that future pandemics are likely, and the COVID-19 illustrates the way a pandemic can change work. This pandemic is likely to have long-term impacts and will serve as a warning for preparing for future ones.
Consequently, the authors of the current review provide a detailed discussion of these and many other factors that may influence the future of work in Supplementary Appendix III. A summary of that discussion is offered in Table 3.

Conclusion
Critical in the future will be how the workplace and work continue to evolve and how the workforce will subsequently be impacted. A preponderance of scenarios and reports indicate the potential for a large prevalence of psychosocial hazards in work or from lack of work. Despite the historical record that technology has generally produced more jobs than it has displaced, concern remains that technology will lead to a future where available jobs are fewer in number and extreme in quality, either very high or very low, with little available middle ground (ILO, 2017a). This is particularly true in countries with higher rates of projected aging. These countries tend to have a larger proportion of older workers at risk of automation (Basu et al., 2018). Two problematic situations are likely to occur. First, many workers will suffer from anxiety and stress due to job possibilities or the lack thereof. Second, although many workers will likely acquire new jobs after displacement, a cohort of workers will not. These individuals will require supportive services and resources that may not be adequately available. To avoid a bleak future, technology and work must be managed collectively in the interest of generating international solutions (ILO, 2017a). Clearly, increased future of work-related global dialogue, support, investment, and commitment will be required with relevant OSH stakeholders and partners at all levels. This review may provide a useful resource for such deliberations.

Funding
Funding for this project was provided by the National Institute for Occupational Safety and Health.