The Knowledge That Walks Out the Door: How Safety-Critical Industries Struggle to Pass On Expertise

When a veteran nuclear power plant operator retires, something leaves the building that no manual can replace. It is the intuitive sense of when a control room reading feels wrong, the muscle memory of a procedure executed under pressure, the instinct honed over decades of shift work. A new systematic literature review published in the journal Heliyon by Rikard Eklund and Anna-Lisa Osvalder of Chalmers University of Technology examines this problem in depth, asking how safety-critical systems—aircraft, ships, offshore rigs, nuclear plants, railways, and control rooms—manage to transfer such tacit knowledge from experts to novices. The answer, drawn from twenty-two studies spanning three decades, is both encouraging and unsettling: the most effective transfer methods are deeply social and informal, yet those very informality carries hidden safety risks that no study has yet examined directly.

The review’s scope is deliberately broad. Eklund and Osvalder defined safety-critical systems as those whose malfunction or failure could result in loss of life, significant property damage, or environmental harm. These are environments engineered to extraordinary reliability standards—catastrophic failure conditions in safety-critical aircraft systems, for example, are designed to be extremely improbable, on the order of one in a billion per flight hour. Yet even the most rigorously engineered systems depend on human operators whose expertise includes a component that resists documentation. The authors searched six databases, including Web of Science, Scopus, PsycInfo, and IEEE Xplore, and screened more than 4,700 records down to a final set of twenty-two studies published between 1993 and 2023, covering nuclear power production, offshore oil and gas, aerospace, maritime operations, railways, and control rooms.

To organize this heterogeneous body of research, the authors turned to the SECI model, the influential framework proposed by Ikujiro Nonaka and Hirotaka Takeuchi in 1995. The model describes knowledge creation as a continuous cycle through four modes: socialization, in which tacit knowledge passes from person to person through shared experience; externalization, in which tacit knowledge is articulated into explicit form; combination, in which explicit knowledge is reorganized into more systematic bodies; and internalization, in which explicit knowledge is absorbed back into tacit understanding through practice. Because the review focused on tacit knowledge, the authors extracted the socialization and externalization dimensions in particular depth, mapping dozens of distinct transfer methods onto this structure.

The socialization findings are striking in their variety. Formal approaches documented in the literature include apprenticeship programs, communities of practice, vocational training, cross-training initiatives, observational learning, and scheduled meetings with experts. Informal methods prove just as important: casual coffee-break discussions, mentoring relationships, free-time social activities, and the family-like dynamics of shift teams, where knowledge is exchanged and reproduced through daily interaction. In the nuclear industry, strategies for an ageing workforce include retiree knowledge retention plans, engaging retirees as consultants or educators, overlapping the departure of experienced staff with the arrival of newcomers, and prioritizing local recruitment to build lasting ties. In the offshore sector, researchers found that managing contract workers—their numbers, timing, and relationships with permanent staff—is a critical practice for preserving knowledge continuity.

Externalization, the attempt to crystallize tacit knowledge into shareable form, relies on a different toolkit. The reviewed studies describe the use of metaphors and analogies, storytelling, exit interviews, brainstorming sessions, debriefings after incidents and near-misses, diaries kept by pilots and operators, think-aloud methodologies in which experts verbalize their reasoning, lessons-learned programs, written instructions, and databases or repositories of accumulated experience. In maritime pilot training, for instance, debriefing sessions and record books capture the judgment calls made during onboard piloting missions. In aerospace, visualization tools have been tested to make air traffic controllers’ task processes and tacit knowledge utilization visible for training purposes. These methods convert fleeting, personal insight into standard operating procedures, best practices, golden rules, and regulations—the formal instruments on which safety-critical operations depend.

Yet the review uncovers a fundamental tension at the heart of this enterprise. Tacit knowledge, as the philosopher Michael Polanyi famously observed, is knowledge we know we have but cannot fully tell. It manifests as practical skills, know-how, muscle memory, intuition, and pattern recognition—categories that overlap and resist clean separation. The authors found that none of the twenty-two studies provided a comprehensive definition or in-depth exploration of what tacit knowledge actually consists of. Worse, research instruments such as questionnaires and interviews may require people to articulate aspects of their knowledge that are inherently difficult or even impossible to express, limiting the development of reliable metrics. What is tacit for one expert may be explicit for another, and the boundary between the two remains stubbornly unclear.

The safety implications of this ambiguity are profound. Informal interaction—coffee breaks, mentoring, social activities—appears to be the primary channel through which tacit knowledge flows. But informal channels can also become undetectable to the organization, leading operators to rely on non-standard procedures instead of established Standard Operating Procedures, thereby increasing risk. In a safety-critical environment, the same social conditions that enable the richest knowledge transfer can quietly erode procedural compliance. Notably, the authors report that none of the reviewed studies explicitly examined the safety risks associated with informal social learning, leaving a significant gap in the literature on one of the most consequential aspects of the problem.

Technology, meanwhile, emerges as a supporting actor rather than a solution. Many studies examined ICT systems, digital portals, databases, simulators, and specialized software as means of retaining and disseminating knowledge. Simulators in particular recur across domains—from air traffic control to train driving to offshore helicopter emergency response—as environments where novices can practice routine and non-routine scenarios under expert guidance. But the review’s consistent finding is that interpersonal interaction remains the recurring theme: technology can support the retention, accessibility, and dissemination of knowledge, yet tacit knowledge transfer appears to depend particularly on interaction between individuals. Organizational strategies such as mentoring, succession training, and retaining experienced personnel as educators complement, rather than substitute for, technological solutions.

The methodological lessons of the review are equally important. Eklund and Osvalder followed PRISMA guidelines, adapted to their context, and applied rigorous critical appraisal tools: the CASP checklist for qualitative studies, the MMAT for mixed-methods research, and the AACODS checklist for grey literature such as reports from the International Atomic Energy Agency. Including grey literature proved essential—excluding it would have substantially reduced the breadth and depth of the evidence, since much of the practical knowledge about workforce ageing and knowledge retention resides in institutional reports rather than peer-reviewed journals. The authors acknowledge residual risks of publication, language, and citation bias, but their multi-database search, citation tracking, and inclusion of non-empirical sources mitigate these concerns.

The review’s conclusion points toward a future research agenda with real stakes. Greater conceptual clarity is needed on the boundaries of tacit knowledge and its relationship to intuition, pattern recognition, know-how, and procedural knowledge—particularly the gap between how practitioners use these terms and how academics define them. Methodologically, the authors call for mixed-methods and longitudinal designs that combine interviews and questionnaires with observations, performance and behavioural measures, and techniques such as eye-tracking, to study tacit knowledge in action rather than relying on self-report. Comparative studies across safety-critical domains and professional cultures could help distinguish context-specific mechanisms from general principles. Until then, the review offers a clear message for industries where failure is not an option: effective knowledge transfer depends less on any single method than on creating conditions in which experience, social interaction, organizational structure, and technology work together—without undermining the safety requirements that make these systems worth protecting in the first place.

Subject of Research: Tacit knowledge transfer methods in safety-critical systems across nuclear, aerospace, maritime, offshore, railway, and control room domains

Article Title: A systematic literature review on knowledge transfer in safety-critical systems: focus on tacit knowledge

Article References: Eklund, R., & Osvalder, A.-L. (2026). A systematic literature review on knowledge transfer in safety-critical systems: focus on tacit knowledge. Heliyon, 12(15), Article e45530. https://doi.org/10.1016/j.heliyon.2026.e45530

Image Credits: AI Generated

DOI: Not provided

Keywords: tacit knowledge, knowledge transfer, safety-critical systems, knowledge management, SECI model, systematic literature review, nuclear power, aviation, maritime, offshore industry, communities of practice, human factors