Assessing artefacts of maritime safety culture—current state and prerequisites for improvement

Implementing safety management (ISM Code) aims to promote a good safety culture in the maritime industry. However, although this has improved safety, it has also paradoxically increased bureaucracy and overlooked operative personnel. At the same time, safety science has undergone a paradigm shift from Safety-I, which is traditional and error based, to Safety-II, which focuses on the potential of the human element. To determine whether Safety-I is dominant in the prevailing culture, or whether any Safety-II ideas are emerging, we studied the current thinking and the prerequisites for improving maritime safety culture and safety management. We interviewed 17 operative employees and 12 safety and unit managers ( n = 29), both individually and in groups, in eight Finnish maritime organisations representing the maritime system (shipping companies, authorities, vessel traffic service, association). We also analysed 21 inspection documents to capture practical safety defects. To the employees, safety culture meant openness and well-functioning, safe work. However, this was not always the case in practice. Safety management procedures were portrayed as mainly technical/authority focused, and as neglecting the human element, such as the participation of operative personnel in safety improvements. We also found several factors that support improving maritime safety culture. The ISM code seems to have supported traditional methods of safety management (Safety-I), but not the creation of a positive safety culture, which the Safety-II paradigm has highlighted. System-wide safety improvement, participation of the operative personnel and open sharing of safety data were the areas in need of development. These were already being created in and among maritime organisations. To improve maritime safety culture in a concrete way and to achieve Safety-II in practice, we need a new focus and competence. Policy, procedures, and practical tools and models should use the human element as potential.