Friday 5 July 2024

Why We Take “Government Advice” - But Shouldn’t

I swear if I read the phrase “according to experts” I will write a letter to the editor of the Daily Telegraph suggesting that they ban the phrase, and substitute instead the name, rank, recent relevant publications, and commercial or State affiliations of the “expert”. Something like
Dr Misha Andry (54), Lecturer in Public Health at the University of Carlisle, a subscriber to the Guardian, a member of Greenpeace and of the National Trust, whose most recent publications have been on the transmission of sexual diseases in gerbils.
This sort of thing should be willingly provided by the “expert” and is often available on WIkipedia or LinkedIn.

Okay. Rant over.

There is a serious side to this. The concept of expertise was philosophicalised (which is now a word, meaning, made the subject of a philosophical investigation or enquiry) by Hubert Dreyfus in a series of essays, in which he laid out a five-level characterisation of skilfulness at a task. It has since been abused beyond belief by HR departments and Training Consultants.

Dreyfus was arguing that the so-called “expert systems” (the ‘computers will replace knowledge workers’ hype of the time) could never replicate the actual decision-making of human experts, becuase truly expert decision-making was highly contextual, used implicit knowledge (in the sense of Polyani’s book with the same title), and could not be distilled into rules.

As a description of how experienced, knowledgeable, organisationally-senior doctors made decisions, he was right. Where he was wrong was assuming that they made better decisions because of it. What little research there is suggests that younger doctors, who are nearer to their up-to-date training, make better decisions than more experienced doctors who have not kept up with the research.

He was, I suspect, over-impressed, as many were at the time, by the confidence of senior medical people. Dreyfus formulated his ideas in the late 1970’s (published in 1980), since when public expectations have risen to the point where today, too many of us have too many examples of friends and colleagues being mis-diagnosed, ignored, and given the wrong treatments and drugs. The NHS has become notorious for its hounding of whistleblowers, and also spending millions on NDAs. We can assume that, if there ever was a time when Consultants were diagnostic giants striding the wards, it is well past.

So the “implicit expertise” Dreyfus described is a myth, but the manner of making decisions he describes as “expert” is surely still with us. I would simply remove “Expert”, with its unavoidable overtones of superior diagnostic performance, from the Dreyfus classification, and add a note to the “Proficient” description that, with time, much of the decision-making and task performance will become more nuanced, seem to be more case-by-case, and almost unconscious. In addition, however, those people have about the same success rate (and its variation over time) as people who make their decisions more consciously.

Dreyfus’ point about limits to the development of AI / Expert Systems still stands. What does not stand is his implict praise for the “Expert” way of practicing.

The lesson of the last twenty or so years, in every profession from banking to public health, is that experts are fallible, and sometimes more-than-fallible. The usual solution is to introduce regulatory guidelines, which will amongst other things, require decisions to be made in a transparent, systematic way that may in addition incorporate compliance with purely political considerations, such as equality legislation. In addition, the lawyers will prefer some kind of documentary proof that these regulatory requirements were followed. This, of course, imposes a bureaucratic overhead of work on the productive staff.

In a technically and legally complex economy, no one person can ever become well-informed enough to take responsibility for all of their decisions. Just as we cannot test every egg for salmonella (nor could we afford the equipment), we cannot understand every tax law, nor everything to do with the working of our cars, let alone anything to do with medical treatment. We have to be able to “take someone’s word for it” and not then become liable for doing so - as long as it is the right kind of person. The principle we need is that competently following “expert advice” absolves one from liability if something goes wrong as the result of the competent application of that advice.

The law define who the “experts” are, whose word we may take on trust. For cars, that is a manufacturer-certified mechanic. For food, it is a licensed retailer. For medical purposes, a GP or Consultant. Within a company, it will be one’s manager, as within the Armed Forces it is one’s superior officer. These people do not absolve one of liability because they are right, but because the law or institutional practice says they do.

On this principle, “Government advice” is not taken because it is right, but because being able to prove that one followed it competently is an absolute defence.

It would be nice if “government advice” was given because it was right, or at least based on the best available evidence and thinking, but of course it is almost invariably wrong. It must be politically acceptable, within the abilities and pockets of most of the population, does not commit Government to spending more money, and is seen to be coming from the “right” sources, which will be members of the Establishment, or, and this is where the trouble starts, “experts” whom the Civil Service are prepared to listen to. The resulting compromises and ideological influences, as well as industry and single-issue group lobbying, almost guarantee bad decisions.

Who should the Government be listening to? One would think, to the people who know most about the issue, who have studied its past and how other countries have dealt with it. Who have recently conducted research about it, and whose papers are cited with respect by other researchers. And whose computer models produce the same answer twice in a row given the same inputs (rather than professors at Imperial College). The “experts”. Who can be cited by Ministers, who are therefore absolved from responsibility. Because they followed the “experts”.

Here’s the Catch-22. If there were a small range of solutions to an problem, that could be packaged up and made available to the public at an affordable price, which would happen if this issue recurred frequently and affected a large number of people, then… well, the private sector would be selling those solutions already and the Government would not need to get involved.

The Government gets involved when the issue is new, infrequently occurring, has a horrendous cost, and there is limited experience and research to draw on. This will mean that there are many competing theories about the causes and remedies, and no way to decide quickly between them. Advice in these cases will require significant technical understanding to evaluate. This is what the Chief Medical and Scientific Officers (and their staffs) were set up to do, but the world has become far too complicated for that to work. And often in these cases, the best lack all conviction, while the worst are full of passionate intensity.

Sadly, in these circumstances, what Governments need is certainty. Even if it comes from people who have produced appallingly inaccurate forecasts before, and whose social and political agendas are barely hidden under their shirts.

And this is how activists capture Government. Not by attacking Whitehall and Westminster with guns and ammunition, but with policy papers, advice, “research” by fellow-travelling academics, and PR campaigns, that are as certain as they are false, and passionate as they are obsessive.

Hence the need to identify the anonymous “experts” who make eye-catching claims with immediate political relevance.

Which is where we came in.

No comments:

Post a Comment