Human Factors in highly automated systems (CIEHF, 25 April 2022)

Summary

This white paper is intended for non-specialists who may have little or no professional background in human factors and ergonomics but who are influential in the way decisions are made about the development and use of technology. The knowledge and guidance it contains is based on both fundamental scientific and applied research, as well as from deep study and learning from adverse events.

The paper is based around nine principles that provide an easy-to-follow guide to human factors issues which need to be addressed when developing and implementing highly automated systems.

Content

The white paper is based around nine principles. Their aim is to provide an easy-to-follow guide to human factors issues which need to be addressed when developing and implementing highly automated systems. The principles are:

1. Understand the potential influence of other elements of the system on the automated components, as well as how the introduction of automation can affect those components. Automation must be seen in the context of the overall socio-technical system it exists in.
2. Recognise that automation nearly always changes, rather than removes, the role of people in a system. Those changes are often unintended and unanticipated. They can make the tasks people need to perform more difficult and can disrupt established relationships, lines of communication and the ability to exert authority.
3. Be clear about which of the four core functions (acquiring information, extracting meaning from it, making decisions and taking action) automation will have the ability to perform for each system task, and under what conditions it will be given the authority to control those functions without human oversight.
4. Be realistic in acknowledging that people, at some level, are going to have to monitor, supervise, and hold responsibility for, the performance of the automation. Design, introduce and support the automation such that those people can maintain awareness of the state of both the automation and the world it operates in.
5. Ensure effective, transparent and unambiguous communication between the automation and the human elements of the system, such that the human is able to remain in the loop and situationally aware at all times.
6. For each task or function an automated system has the ability to perform be as explicit as possible where the balance between authority, responsibility and control lies. Be clear about what the expectations about responsibility imply for the different stakeholders in the system.
7. Ensure the people relied on to support the automation understand what the system is doing and why. There should be no automation surprises.
8. Avoid making unrealistic assumptions about the ability of people to monitor and effectively intervene in any system where there is little for them to do over sustained periods.
9. Recognise that automated systems can increase the levels of task difficulty and workload imposed on the human elements in the system as well as the level of human reliability needed in the inspection, calibration, maintenance and testing of system components.