Keywords

1.1 Safety Today: A Dominant Model

Describing how safety works in high-risk industries, and more specifically the role played by humans at individual, organisational or even societal levels can be done in different ways depending on the level of detail and specificity one decides to adopt. From a very global perspective, one could say that the “visible” safety frameworks in high-risk industries share similar characteristics. Yet, when going into further detail, closer to the field and real practices, one needs to introduce some nuances to the macro-“exhibited” model of how safety is ensured.

This section starts with an overview of the macro-safety model that underpins safety in high-risk industries, or more precisely, the way industries are expected by society, and thus regulators, to ensure the safety of their operations. It summarises the main characteristics of this global safety “as demonstrated” view. It then zooms in closer to the field and underlines some discrepancies between this “control” perspective and the actual practices contributing to the safety of high-risk industries, including non-proceduralised ones.

1.1.1 Main Characteristics of the Dominant Safety Model in High-Risk Industries Seen from a Distance: Formal Similarities

Historically in high-risk industries, the dominant safety model in terms of the role of humans has been based on the compliance with rules, processes and procedures at all levels, from governments through laws, to first-line operators through detailed procedures [7]. Hazardous industries are highly regulated from the outside with safety authorities developing regulatory material and ensuring oversight. This approach reflects the societal expectations of total control and clear responsibilities for the safety of new technologies with high damage potential, through the reduction, if not elimination, of uncertainties. For similar reasons, high-risk organisations are also highly proceduralised from the inside with organisational processes and procedures to standardise practices at all levels and avoid variability.

Anticipating everything, developing a priori safe responses and expecting humans to implement them is the driving force of safety as imagined and demonstrated. The underlying philosophy is to reduce uncertainty to almost zero and control risks. In such perspective, uncertainty comes down to epistemic uncertainty. Any unanticipated situation is explained through a shortfall in the risk model which is completed after the facts and still believed complete (or complete enough).

Since the 1990s, Safety Management Systems have been progressively adopted in many high-risk industries with the ambition to move away from a compliance-based to a performance-based approach to safety, whereby experience is used as a feedback loop to improve risk models, and organisations have more leeway regarding the risk reduction measures and are able to identify those best suited to their specific characteristics. This evolution involves a clear definition of safety accountabilities. A major characteristic of safety management in high-risk industries has always been that it ought to be auditable (whether from the outside, the inside or both) and accountabilities should be easy to establish.

This change of scale from a regulatory authority defining all the standards to be complied with to organisations taking a more important role in it and integrating the lessons from experience remains in line with the foundations of the previous approach and societal expectations about the safety of high-risk industries.Footnote 1 Overall, the safety model is still based on the belief that uncertainty can be reduced to almost nothing if not completely eliminated and that safety can be totally controlled this way. In such a framework, the role of humans is preconditioned.

1.1.2 A Closer Look at the Dominant Safety Model in High-Risk Industries: Some Nuances

Although this global framework of safety as imagined and as demonstrated has its own internal consistency, practices are not fully in line with it as already extensively documented in the scientific literature.

Within industries, although most of them rely on SMSs, the content and what it corresponds to varies significantly from one organisation to another, as does the credit they accord the usefulness of the SMS. Practices in terms of SMS implementation vary from mere compliance with the SMS regulation seen as paperwork to an actual reflexive activity of how to enhance the way safety is managed that most often also entails other activities beyond the SMS itself [3, 4].

In the first case, safety is managed through the demonstration model only with clear accountabilities, roles and procedures supposed to be sufficient to control risks. A number of high-risk organisations turned to consultants to develop their formal SMS, due to a lack of internal competent resources, thereby opting for another source of standardisation, that of the consulting company [1].

In the second case, organisations acknowledge the inevitable existence of contingencies and develop other means beyond risk control to cope with real situation uncertainties, such as resilience capabilities. However, in parallel, they need to maintain a formal demonstration that everything has been put in place to manage safety according to the auditable SMS requirements. Therefore, two worlds coexist within advanced organisations: (a) a formal safety management one used to provide external demonstration that safety is under control and (b) other safety enhancement strategies supporting a world of practices described in the scientific literature showing for example that: there is more to first-line operator activity than a mere implementation of procedures; operational managers rely on a combination of practices beyond formal processes to actually contribute to safety [2, 8].

Even at the safety governance level, where the formal model as imagined and asserted involves a clear boundary and separation of roles between authorities and industrial organisations, practices are more complex. In order to compensate for the lack of qualitative and quantitative internal resources, regulatory and oversight authorities sometimes involve big industrial companies in the development of regulations and acceptable means of compliance that then serve as a basis for the entire industry.

1.1.3 From a Demo Version Based on Risk Control to Real Versions also Calling on Adaptation

Although the social control of risks in high-risk industries still relies on a formal model of risk control where uncertainty is reduced to almost nothing, practices differ to some extent from this external demonstration framework to actually cope with the inevitable contingencies of the real situations faced at all levels, from first-line operators to higher managerial levels. Yet, some of these practices escape some of the key foundations of today’s social control, especially (almost) total control, auditability, accountability. To what extent will these foundations be challenged or reinforced in the future, and thus possibly affect safety models (either formal or practised or both) and more specifically the role played by humans in the safety of high-risk industries?

1.2 Why Should We Think About Safety Differently in the Future?

Although so far, the evidence of the limitations of the “as demonstrated” safety model has translated into the development of dissonant parallel practices, some major trends are foreseeable in the future in many areas. This section explores some of these trends, especially the technological evolution, the socioeconomic and industrial evolution and the societal and sociopolitical evolution. For each of them, it:

  • briefly describes the main trends, from the ones already underway to more uncertain or speculative ones;

  • analyses the challenges or opportunities they already represent for existing safety models, either “as demonstrated” or as practised;

  • discusses how they could possibly further destabilise or reinforce the current models in the future.

With all the humility that is necessary in such an endeavour, the following subsections address the aforementioned three main areas of evolution and their possible impact on safety models. The trends are addressed independently from one another for the sake of readability, hardly ever touching upon their links. Yet, they are interrelated with one another, with possible emergences that cannot be anticipated.

1.2.1 Technological Evolution, Especially Massive Digitalisation in Many Areas

The digital revolution is already underway in many areas of our personal and professional lives. More and more data are collected in more and more domains, with increasing capabilities to process them and to cross-reference different sources and types of data. The reign of artificial intelligence (AI) is proclaimed for tomorrow if not today with huge impacts announced in many fields. This trend, although massive, is to be nuanced in practice though when considering the scope of applicability of AI.

Nevertheless, several changes at several levels likely to affect safety are made possible by this technological leap. For example, at management level, the boundaries between domains (e.g. production, maintenance and safety) or between professional and personal worlds could be blurred by the crossing of data from different sources. At a more operational level, new technologies (AI) are already leading to new systems and forms of interactions. Although their use in high-risk organisations is still limited, especially for safety critical functions, they could further develop in the future in these areas as well. If we broaden the picture, old and advanced technologies will coexist in work situations or at least in organisations. This situation is not unknown and already started with automation decades ago, but the trend seems to be accelerating and generalising. In this landscape, an evolution of jobs and required competencies is anticipated, even though real figures are for the time being still lower than the storytelling around AI is suggesting [9].

At a wider societal scale, authors like Matyjasik and Guenoun [6] envisage a possible shift from today’s governance regime to other regimes enabled or fostered by these new possibilities, even though the symptoms of such evolution remain timid today.

1.2.1.1 Main Challenges Faced by the Current Model in the Face of this Technological Evolution

The massive use of digital technology has already started to challenge the dominant safety model in several respects.

One of the areas most challenged by artificial intelligence is the certification of technology. In the “safety as demonstrated” model, certification plays a major role, whereby a new technology is considered safe from a priori safety analyses, even though they are refined and completed through experience. Certification of technology is currently based on the assumption that the technology behaves as it did at the time of certification. This assumption can no longer be made with AI and its evolving performance by design.

The massive digitalisation trend also leads to an increasing exposure to cyber-threats, some of which are possibly affecting safety. This calls for considering safety and security as intertwined, although they have historically been addressed separately from one another and still are today in terms of both governance, management, methods and research.

1.2.1.2 Other Possible Technology-Related Sources of Destabilisation of the Current Dominant Safety Model

Beyond the challenges that the current safety model is already facing, other aspects of the massive digitalisation trend may contribute to changing the safety management landscape in the future, although with great caution from the authors considering the level of uncertainty.

The increasing use of and value attached to data may create an imbalance between those who have the data and the capabilities including competencies to process them and those who do not. This may lead to new roles between safety regulators and industrials. Indeed, the data regarding operations are mainly gathered by the industrial partners themselves and the capabilities to process them are currently developed internally as well to support the overall organisational performance. An increased reliance on data to support safety management and safety governance puts safety authorities in a position where access to data and associated competences to process them needs to be negotiated with industrial companies.

The current governance approach could be even more destabilised if, pushed by new technological capabilities, authorities actually stop working in silos. This would lead to an evolution of structures and means of oversight, and possibly, as envisaged by Matyjasik and Guenoun [6], to an evolution of the actors of governance.

The uncertainty is huge as to whether the importance given to data will continue to grow or the use of data will be framed one way or another in the future. Even though one can only speculate at this stage, the massive use of data raises some additional questions as to how it could interplay with safety in the future, such as: could it possibly lead to the merging of data gathered in the professional scope and personal data gathered outside? Could safety monitoring use personal data on top of professional information?

Regarding the possible influence of digitalisation on the philosophy of the current dominant safety model, views diverge. Some believe that massive digitalisation can increase the power to anticipate (e.g. through numerical twins) and thus further reduce uncertainty, which would reinforce the historical philosophy of safety. Others believe on the contrary that the massive digitalisation will increase uncertainty, and thus the possibility to be surprised (a kind of new normal digital accidents). This rather calls for a different role of humans in the safety of high-risk industries, closer to that observed in practice but not officially and socially acknowledged. Whether it would change the view on social control of risk is still an open question.

1.2.2 Socioeconomic and Industrial Evolution

Beyond technologies, other evolutions are ongoing or foreseen in the socioeconomic and industrial areas.

Demographics is one of the less uncertain trends. Crossed with the socioeconomic dimension, it may lead to the ageing of professional staff, as has already started in Japan for example. This means a wide span of ages, each with their own competencies, motivations and perspectives, working in the same company and sometimes together.

Regarding industry, the trend that started in the 1980s with the emergence of fragmented and interconnected organisations continues to develop at an international scale, including in high-risk industries. It means that several sites, countries and nationalities are involved in making the organisation work. Although this trend is not new per se, what might change in the future is the distribution of activities across countries. Whereas major international corporations used to be led by western countries, the distribution of industrial power has started to evolve. The COVID-19 pandemic might have led some countries to considering the relocation of some industries to Europe, for example, but whether this inclination will actually turn into fact is still uncertain.

What seems to be increasing is the power of the major economic actors, especially compared to that of national states and their regulatory authorities. This phenomenon leads in some cases to what some authors call the capture of regulatory authorities, for example through an influence on the recruitment of authorities’ staff, and thus challenges the independence of authorities, one of the current pillars of safety governance.

In the socioeconomic area, an “uberisation” trend of the economy and labour relationships has already started in some countries, with massive externalisation and subcontracting. This evolution leads to more imbalance between employers and employees through the weakening of unions and an increased power of professional elitism in the hands of the happy few in a context of scarce employment. To what extent could this trend be generalised at the international scale remains an open question.

1.2.2.1 Main Challenges Faced by the Current Model in the Face of this Socioeconomic and Industrial Evolution

The socioeconomic and industrial evolution reinforces and complicates some questions that have been around for a while, especially among academics but that the dominant model managed to escape so far. However, by making some challenges more obvious and salient, it could reach a point where even the dominant model would be destabilised.

The acceleration and generalisation of fragmentation and interconnection of organisations as well as subcontracting and externalisation challenges the development of safety culture. Safety culture is needed not only within organisations, but also across organisations that have different objectives in a context where the sense of belonging to an organisation, or even to an industry, is low and the labour relationships decline including as far as safety is concerned. The ageing of professional staff may add to this diversity of profiles having to work together, and thus to the challenge of developing or maintaining a safety culture. Although safety culture is not always directly or clearly part of the dominant model in the sense of a dedicated regulation and oversight of this aspect, at least the narrative around safety culture might have to be reconsidered.

Another direct challenge to the dominant model is posed by the socioeconomic and industrial evolution: the one-size-fits-all requirements (whether regulatory or internal) and safety management model focused on individual organisations. With the evolution underway and accelerating, the very notion of organisation is destabilised. Furthermore, these requirements are increasingly recognised, including within big international industrial companies, as not being applicable everywhere since they are decontextualised.

1.2.2.2 Possible Other Sources of Destabilisation Related to Socioeconomic and Industrial Evolution of the Current Dominant Safety Model

Today, the regulatory requirements are developed by major Western industrial actors supporting regulatory authorities, and then imposed on the industry as a whole, that is across organisations of different sizes and natures and across the world for internationally “regulated” industries such as aviation. Although, there is currently some interest for both major industrials and regulators in this approach, at least around the “safety as demonstrated” model, the main influencers could change in the future with major industrials possibly changing hands and continents (e.g. from Europe or North America to Asia) and regulators from non-western countries becoming more influential. Would new standards be developed? Would the whole model of regulation and oversight as it exists today be revisited? If yes, to what extent? Would high-risk industrial companies decide to install their headquarters in countries where safety regulations are less stringent? Would they rather intervene in the recruitment of regulatory staff and control it to a certain extent, thereby challenging one of the key conditions of the safety as demonstrated model, that is the independence of authorities? Although uncertain, these possible evolutions could destabilise not only the predominance of western standards, but also possibly the safety governance regimes and the “safety as demonstrated” model that so far has provided a kind of illusion of control that still seems socially acceptable, that is sufficiently reassuring at least in appearance to all safety stakeholders (see Fukushima).

1.2.3 Societal and Sociopolitical Evolution

Some trends are also observed at the societal and sociopolitical levels that could also impact existing safety models. For example, although trust in public services was already challenged in the 1980s, the scientific controversies around climate change or even more recently COVID-19 seem to have further increased the distrust towards experts and institutions. At the same time, an increasing attention is paid, including by civil society, to concerns like climate change or health care. In some cases, this attention also translates into a growing involvement of civil society (or a desire to be involved) in the governance of topics citizens feel concerned with.

All this is happening in a context where access to data and information (whether validated or not) is much easier and faster than it used to be. Today’s communication environment (e.g. social media) enables a fast and broad dissemination of information, and thus societal mobilisation.

On another note, with the COVID-19 crisis, the interrelations between various stakes (e.g. health, economic, social) have been extensively illustrated. Although this is neither new nor limited to healthcare crises, the acknowledgement that multiple stakes coexist at the same time (likewise, high-risk industry means employment, economic activity, etc., at the same time as safety) may become more explicit and challenge the current model of governance, especially that submitted to external demonstration, addressing each stake individually.

1.2.3.1 Main Challenges Faced by the Current Model in the Face of this Societal Evolution

The current safety governance model involving industrials and regulators, who are supposedly independent and representing the voice of the public, was challenged in several cases such as following the Fukushima nuclear accident or more recently the Boeing 737 MAX crashes. In the latter case, the investigation explicitly and publicly pointed out the lack of independence of the US Federal Aviation Administration in its oversight duty. Although it is not the first case where blurred relationships between industrials and regulators were highlighted, the current information and societal environment might amplify the impact on the social acceptability of the current governance model from which civil society as such is absent.

1.2.3.2 Possible Other Sources of Destabilisation Related to Societal Evolution of the Current Dominant Safety Model

More generally, building on what happened with climate change, where the societal pressure increased to a point where it pushed the political sphere to evolve and take this societal concern more seriously than before, one could anticipate an evolution of the current safety governance model, and thus of the “safety as demonstrated” model. Whereas today, this model mainly involves industrial companies and regulators, the increasing external scrutiny carried out by civil society could turn into more acknowledged and formal roles in the future.

In this power game, the control of information becomes crucial. Combined with the digitalisation mentioned earlier, the forces might be imbalanced between the industry collecting more and more data and regulators and the civil society possibly more dependent upon the data and their processing, and thus interpretation.

Furthermore, in times of growing public interest in health and climate change issues, the importance given by the public society to safety could evolve and a new hierarchy emerge among all these aspects.

1.3 Could the Models of Safety as Demonstrated and as Practised Converge?

Overall, the identified trends could destabilise in several ways the current dual approach to safety, where the model of safety as demonstrated and as practised coexist although they are widely acknowledged to be different. Whether the challenges and opportunities these trends represent will increase the gap between the two, possibly up to a crisis of the “safety as demonstrated” model, or will conversely reduce the gap, cannot be predicted. Yet, from different angles, both new societal expectations and new tool capabilities are likely to drive changes in the way safety is managed in the future.

1.4 Book Outline

The following chapters expand on some of the evolutions mentioned above. Although not addressing all of them, they provide different perspectives on some of the major changes likely to impact not only safety as demonstrated but also safety as practised.

It first explores the issue of digitalisation and the many challenges it involves.

In Chap. 2, De Boisboissel takes us to the military world where more and more robots and artificial intelligence are being introduced in war equipment and vehicles and operations are increasingly commanded remotely from the battlefield. The author explores the questions raised by this evolution, especially regarding responsibilities, decision-making and ethics. Based on these insights from the military, Laroche and Reuzeau derive, in Chap. 3, six main takeaway points that civilian organisations could reflect upon.

In Chap. 4, Allspaw takes us behind the scenes of digital devices, more specifically of critical digital services, Internet-connected services that have taken on greater importance in the functioning of society. The author proposes an original perspective on this recent domain by looking at it as a safety–critical field and presents the challenges but also opportunities and advantages that it offers.

Stian Antonsen, in Chap. 5, somehow moderates the enthusiasm towards digitalisation. The author comes back to the scope, capabilities and biases of artificial intelligence, especially compared to human intelligence, and puts them into perspective with the expectations in terms of safety from high-risk industries. He highlights a number of paradoxes that will need to be addressed before AI is used in safety–critical decision-making.

Moving to the impacts of socioeconomic evolution on safety, a number of chapters reflect upon the challenges posed, especially to safety, by the demographic evolution in terms of career length and competence aspects.

Chapter 6, authored by Tosé and Tazi, presents the current situation in Japan where the retirement age was recently pushed back to 70 and will be pushed even further, to 75, in the near future. What Japan is facing today could give a flavour of what other countries could experience in the coming decade and thereby could serve as a source of inspiration. The authors address some of the challenges such as skills and knowledge updating, keeping an interest in work for all generations or managing the mix of generations, which could in turn impact safety. As they envisage to turn senior people into trainers, in Chap. 7, Navarro Rodriguez and Largier highlight the conditions needed to support this mentoring activity for seniors, based on research work done in France and more specifically in the nuclear industry.

The two following Chaps. 8 and 9 put a special emphasis on competencies. In Chap. 8, Pons, Rodriguez and Reuzeau share the work launched by Airbus in 2018 to “meet the future competence challenge” as emphasised by the chapter’s title. In Chap. 9, Largier comes back to what competences entail. He underlines the importance of reaching beyond a purely managerial approach and considers organisational dimensions to account for how competencies are actually mobilised and developed individually but also collectively in work situations.

Barcellini, in Chap. 10, reflects upon the management of work transformation when introducing new technologies to avoid repeating pitfalls experienced in the past. She suggests considering the design of future work as a transition process that involves several dimensions—social, organisational, technological—that all need to be addressed to shape a “future of work”.

The societal and sociopolitical evolution is touched upon by Baram and Bieder in Chap. 11, through the prism of regulation of high-risk industries. The authors explain how the self-regulation initiated a few decades ago is in the process of being pushed to the extreme and leading to the subordination of safety regulation to industrial leadership on risk governance issues, setting back the role of society on these matters.

The evolutions addressed in the various chapters and referred to earlier in this chapter are not independent from one another, making the overall context a complex and dynamic landscape in which unique human capabilities may well be essential for keeping high-risk industries safe. This is what Shorrock reminds us in Chap. 12, drawing seven learning points from clinicians testimonies on their experience of working under rapidly changing conditions during the COVID-19 pandemic.

Lastly, Chap. 13 by Pariès attempts to draw some lines of thought and action on the management of future safety challenges from the insights provided by the previous chapters. It especially reflects upon the gap between the levels at which major changes shaping the future of high-risk industries are decided and occur and that at which safety is discussed.