My Lords, I will make a comment on Amendment 1, but perhaps also a more general comment on some of the amendments in this group and the next. I absolutely recognise that these automated driving systems need to be accepted by the public, and I see that many of the amendments look to do that by increasing the emphasis on ensuring safety. I am sure that we are all hugely sympathetic to that, so I am very sympathetic to the motivation for many of the amendments—but we need to be a bit careful. This is a new area of technology—it is still developing, of course—and there may be some unintended consequences of some of the changes that people are proposing.
In talking about this, I will briefly refer to some recent research by Konstantinos Mattas and his colleagues from the European Joint Research Centre in Ispra in Italy. For example, in Amendment 1 we are asked to include a phrase about
“substantial real testing on roads”.
I think we really need to explore the value of this. Human drivers have a frequency of fatal crashes of one every 3.4 million hours, which would imply continuous driving for 380 years. To demonstrate that automated vehicles are safer than humans, you would need to run 100 automated vehicles for 24 hours a day for 225 years. Of course, if you change the software or the hardware, you might well need to repeat tests of this length. We need to recognise the challenge of demonstrating safety by long periods of testing on roads. The vehicles will have to be significantly unsafe for a realistic period of testing to start to show up the problems.
I cannot support Amendment 1, even though I hugely sympathise with it, because it is an apparently simple ask but I do not think it is likely to deliver the benefits intended. Unfortunately, it could be counterproductive, so I think the wording of the Bill as it stands is preferable to the proposal in Amendment 1.