In recent years the development of autonomous vehicles has increased tremendously and a variety of methodologies had been applied to make them more safe and secure. This work shows a multilevel approach combining Failure Mode, Effects and Criticality Analysis of an autonomous railway system with sociological and technical aspects to support safe operations and human-machine interactions in the field of autonomous railway systems. This approach includes all relevant technical components, as well as the assessment of measures for a safety process based on the Failure Mode, Effects and Criticality Analysis. We applied the Persona-Roberta model to assess safety aspects at the interface between humans and machines and applied both results to establish training materials. The results provide answers to questions about the avoidance of technical errors, discussions on security and safety aspects and shows organizational development tools for accident prevention. In the future the created knowledge will be used to improve trust in digital solutions and Cyber-Physical Systems.