David Slater June 6, 2020
These days there seems to be a false dichotomy between new and old thinking in our approach to ensuring safe environments for both public and professionals. The reasons have been presented ad nauseam by both “sides” in this argument. It goes something like this.
Traditionally, the emphasis has been on preventing failures, with barriers and systems to defend against disruption of intended (as designed) operation. This could be by deliberate, or accidental interference, or performance aberration, by human, or physical agents. Another view is that “normal” operations of systems are always being disrupted by the natural variabilities of real life situations. “Normally” we are able to cope and the systems work, but sometimes (occasionally), the deviations can overstep the limits that the system can adapt to. Here, properties of the system and the operators inherently tolerate / adjust to allow normal operation and can be developed (learned / acknowledged) to increase the limits of successful (safe?) operation.
Presented adversarially, in the first mind-set, humans are a problem in that they are the least predictable / reliable components in the system. In the second mind-set, successful (safe) operation relies on the resilience of the human operators to work around the problems.
But standing back from the PR, isn’t it obvious that, particularly as systems get more and more complex, we need to design systems to be as reliable as possible? But the world does not always obey the rules and stick to the standards! So we do need to recognise that we should also design in or implicitly rely on some extra resilience in the system as insurance. So it’s not either/ or, SAFETY I or SAFETY II, it’s both! Its belt and braces (to paraphrase Kletz).
There are of course recognised, tried and tested ways of ensuring the inherent safety of the designed system. What is not so well established is how you ensure, or design in the bracing resilience; and what is it in the human behaviour that naturally provides this extra dimension?
Current thinking on designing in resilience, looks to add and ensure the system has functions that allow it to not just respond to and monitor the effects of, disruptive events, but formally to learn from patterns and the results of upsets and to be able to anticipate and take pre-emptive action to mitigate or offset expected outcomes. This can be represented as the FRAM diagram below.
But if we look at how the human mind works – how we think, we see that this is exactly how the brain has evolved and is structured to make decisions, in uncertain environments. It seems to have evolved as an increasingly complex and sophisticated series of layers.
There is essentially a primitive brain around the spinal cord and the brain stem which is hard wired to automatically (autonomically), take care of breathing, digesting, circulation and reflex responses to an array of sensors providing stimuli.
There is then the next level of functions (the Thalamus and Hypothalamus) that monitors and sorts these sensory signals and routes them to the right displays and provides the necessary resources needed for response. A third function the Cerebellum has all the programmed circuits which provide planned actions and necessary skills)
The next level sometimes called the mid Brain, then provides through the Hippocampus the ability to store and recall events and patterns as learned memories. Finally the top layer in the brain is the cerebral cortex where all this sensory data is organised, compared against experience and predicts what should happen next: correcting continuously prior guesses with updated data. Yes the brain is Bayesian! We can again use a FRAM picture to show this below.
So if we compare the two, we can see that the human brain is a natural resilience system, which we’ve been adding to our designed systems since man started using tools. So in a way this is a false dichotomy and the only thing we have to realise is that the human has the ability to add that extra insurance naturally and therefore, from here on in we shouldn’t be arguing either / or, we should be looking to optimise the human factors in the system interfaces (Human Factors). So let’s stop arguing. Resilience is natural, it’s essential. We’ve always had it – let’s use it intelligently?