Discretion and safety

Here in the UK we have, this week, seen a minor political and family tragedy played out in the Vicky PryceChris Huhne case.

English: Chris Huhne at the Liberal Democrat S...

In short, Huhne, appointed in May 2010 as a Liberal Democrat member of the Cabinet, left his wife, Pryce, for another woman in 2011. She then sought revenge by telling journalists that, around a decade ago, she had lied to the police on his behalf to take the blame for a speeding incident. It went to court and Huhne pleaded guilty to perverting the course of justice and resigned as an MP (he had quit the Cabinet when charged) and last week Pryce, who had sought to rely on the archaic defence of “marital coercion” was convicted of the same offence. They both face the likelihood of jail when sentenced.

I knew Huhne a little through my work with Britain in Europe and I am sad to see his career ruined in this way, but that’s not what I am going to write about.

Because what really ruined Huhne was his decision to break the speed limit. If he had not been able to do that then he would still be a Cabinet minister even now. And it is realistic to believe that we will be able to stop people making such decisions in the next decade or so – if we want to.

At first glance it seems obvious that we should. Speeding cars kill people – rather more serious a matter than a domestic and political tragedy – and stopping them would save lives, probably disproportionally amongst the young (as drivers, passengers and pedestrians) and the poor.

The technological potential will certainly exist in a decade’s time. Today’s cars are already filled with microprocessors and the ongoing shift in chip fabrication to multicore systems means that cheap and reliable safety critical systems will be possible. Such systems do not just need to control a car’s speed – they also need to be able to survive component failure, and the response to the problem of the “power wall” in chip manufacture – build lots of lower powered chips into devices – means that the necessary redundancy will likely become available.

But will we do it? I have a bad feeling that we won’t.

Motorists are one of the most powerful lobbies in western societies. They tend to be wealthier and better educated than pedestrians and they also have an ideological pull in the debate where “libertarianism” has more supporters in the political class than it does in the real world (ie., politicians are more political than ordinary people and so see these debates in ideological terms). The freedom of “white van man” to drive like a maniac will turn into some “save Britain” issue for some people, I am sure.

The most comparable issue to me seems to be doctors. Doctors kill people, not deliberately, but because we give them too much discretion. If we sent more doctors on to hospital wards with computers that told them what to do instead of relying on doctors’ professional knowledge then we would almost certainly treat more people in a better way. Of course, we do not need to send doctors out at all in those circumstances – we could send nurses or some other forms of “para doctor” to do the job. These people would have enough medical training to be aware if the machine was malfunctioning and to carry out its instructions properly; but they would not require seven years at University medical school.

And the money we saved could be invested in better training for the doctors we did have, better public health, better drugs or better public parks – whatever.

But, again, doctors are a powerful constituency and quite willing and able to scare people with inaccurate stories about evil computers and the like: only the bravest of politicians takes them on – and even then not for too long.

This failure to control discretion is going to become an ever bigger elephant in the living room in many aspects of our lives in future. We have already seen how the internet has been smashing deference to medical professionals by making amateur analysis ever more powerful (I do not pretend this is not without its problems). But the revolution will become more intense when it enters the world of the professionals themselves. As another example think of lawyers, do we really need them for tasks like property conveyancing if we can build expert systems that can both manage a conveyancing algorithm and access the public data needed to ensure due diligence is applied?