Traditional approaches to safety will not succeed in further reducing the now very low accident rate in commercial aviation. A radically new approach is needed, argued Griffiths University Professor Sidney Dekker at the First South African Symposium on Human Factors in Aviation, in Boksburg, east of Johannesburg, on Tuesday.

The traditional approach to aviation safety sees people as the problem -- as shown by the prevalence of phrases such as 'human error' or 'pilot error' -- and measures success in terms of an "absence of negatives". There is seen to be a correct way to do things and this must always be adhered to. All that is required is to ensure compliance. Accidents are the result of non-compliance, which can be punished.

"There is," however, "always a gap between how work is written up and how work is done," he cautioned. "This gap is filled, not by violations, but by judgement, expertise." All too often, you cannot follow the set procedures and get the work done. "Stop looking at where things go wrong. Understand how people make things work, despite all the things that conspire against them."

"In 2014, we have become victims of our own success," he affirmed. "Our [aviation safety] success is absolutely epic. We are a beacon for other worlds; the oil industry, the medical industry, look at us." But although the accident rate has dropped very low, it is not yet zero. Plotted on a graph, it shows a narrow gap above the axis. Dekker referred to this as "the thin edge of the wedge".

"Doing more of the same won't help," he said. "Doing more of the same means you get more of the same. The thin edge of the wedge means our strategies are failing." The key point is that, in the early twenty-first century, "accidents are the result of drifting into failure". Incremental developments within an organisation, each safe in themselves, build up until the cumulative effect causes an accident, which is then blamed on human error and people are prosecuted.

"Murphy's Law [everything than can go wrong will go wrong] is wrong: everything that can go wrong usually goes right, and your organisation draws the wrong conclusions," he summarised. "Then, one day, conditions line up against you and there is an accident."

"The enemy of safety is the fact that your organisation is successful," highlighted Dekker. "The enemy of your safety is your success. Past success in a dynamic world does not guarantee future success." Safety will be dependent on giving personnel the capability of operating safely, creating a situation in which they trust their managers and their managers have confidence in them.

Issues of risk must be discussed when things are going well, and risk discussions must be kept alive when everything seems to be safe. The atmosphere in the organisation (and not just paper directives) must make people feel that they can stop something for reasons of safety. "Let's go into a future in which your people are your most important resource," he concluded.