Robots 'could replace 250,000 UK public sector workers'

There's a huge gap in the market for meaningless cures, I don't know, maybe people could just stick needles in people, or listen to them, or lay on hands, or wave crystals, or crush flowers in front of them and wave the scents, stuff like that. Surprised nobody's doing this kind of thing.

Which is why I am always hostile to the medical profession's "it is no better than a placebo effect"

"Yes, but is it as good as a placebo effect and is it cheaper, including clinician time and cost of premises, than a placebo?"

The NHS should be employing more homeopaths not fewer.

You would be surprised about what the IT sector can do, especially when another sector (i.e. healthcare) is under pressure and considerable stress (i.e. an aging population and tightening budget).

Hovercars were never even close to being invented because they are almost pointless from an economic standpoint. Such technology for healthcare certainly is not, and is likely to be almost crucial with the aging UK.

Although I do likewise think it will take a long while, it is definitely plausible to see such technology within our lifetimes (born 1990-2000+) thanks to the magic the market has on 'stressed' economic sectors and continuous technological improvements.

Up until now there was not the same pressure being placed on healthcare. Not as fast an aging population and more immigrant doctors/nurses without such likely budget cuts. Likewise, the IT sector was nowhere near as developed as it has only properly started snowballing from around the 2000s.

Exactly this, it's not a new trend, and it's one we've clearly survived in the past... ironically the advance in technology is one of the things that seems to be making people worry more despite the benefits it's brought in the past; there's a lot of scaremongering in the news etc, without the positive side of the story being given any notice.

Ok nuli lets try this out.

"is it as good as a placebo"- No, homeopathy is not as good as the most effective placebo.

"and is it cheaper" - No, there are far cheaper ways of administering placebos than homeopathy.

Honestly it simply wouldn't work. Let's pour aside all other roles of a doctor and just look at diagnosing:

Human error in medicine is definitely under-recognised. Most patients have no idea when a serious error has been made and just assume it's normal. In fact, patient's are no more likely to complain about a doctor who had made a mistake than one who has not.

But saying you could currently create an AI capable of objectifying all of the subjectiveness and subtlety of human interaction is not going to work. Your computer can't smell alcohol on someone's breath and then ask about it. It can't quantify someone's pain the same way a human can (everyone in A&E says their pain is '10/10'. It can't listen to a chest. And if you're getting a human to input such clinical values then that's not exactly an independent AI.

You might stand a chance once biochemistry really takes off and you could do 1000 tests per patient to have enough data, but that is 50 years away.

Come off it, a radiologist stares at scans and diagnoses whatever. You could programme a computer to detect masses of particular darknesses, location, density, size or whatever and divide them into three classes-"got the disease", "not got the disease" and "further tests needed".

The computer will not get tired or distracted and will not make the same careless mistakes as the radiologist. However, there will be certain images that don't fit the computer's criteria of "further tests needed" where the radiologist thinks "not happy about that one. We will order the further tests anyway" and some of those will turn out to be the disease and perhaps a few will turn out to be some other health problem that wasn't even under consideration.

Tbf, most machines pre-digital were not really autonomous in the sense that 'robots' are intended to be, they acted as mechanical aids to human work and required constant human supervision. A good example being the sort of calculating machines that prevailed from the 1940s onwards.

I agree. Image analysis is one area that might have promise and its easy to quantify. Although you'd have to replace 'further tests' with 'needs human analysis' - many findings are only interpretable alongside history, examination and lab findings. And even then, we still aren't quite there yet technically, let alone legally and then the time it takes for such things to be 'on the ground'.

But you've picked one small area of medicine. How is a neurologist's diagnostic ability going to be replaced when its so dependent on clinical examination? Realistically the only ay i can see that working is waiting for scan image resolution and biochemical/immunological tests to greatly increase their data so that they can then be used in isolation without a clinical story. We're decades away.

I do not think there are legal implications. The law has rarely stood in the way of automation. The law finds it easy to say an automated process is a reasonable one.

So small parts might be automated yes And as mentioned above - there are some rapid tests for various infectious diseases emerging that could open pregnancy-test style automation for a very limited number of conditions. But not your original statement.

I wouldn't know about the legal side but I was under the impression self-driving cars were having problems with determining liability and that companies were less willing to take even basic steps as a result. I think such a problem in healthcare might be even more pronounced as a result. But maybe not, although certainly the NHS approval process alone will take many years (hte hospital I'm working in now has decided it wants an electronic prescribing system, something pretty simple that has been used for >10 years elsewhere in the world, >5 years in other UK hospitals. The earliest they are able to implement is 18 months away).

I would like to see evidence regarding this

Are you talking about a consultant prescribed placebo or a GP prescribed placebo?

What is the average unit cost of a consultant consultation? What is the average unit cost of a GP consultation? What is the average unit cost of a homeopathic consultation? Those costs are of course not merely the cost of their respective labours, but also their establishment costs. Any consultation in a hospital will have a greater cost than that in alternative therapist's consulting room, simply because hospitals are vastly more complex and expensive buildings per square foot.

Again, the medical profession play a game here. When they cost out placebo therapy they basically price the sugar pills and nothing else. However when they price alternative remedies they include the therapist's time cost.

The NHS won't have the very serious automated car problem because it doesn't matter whether it was the porter who dropped you off his trolley, the nurse who failed to feed you, the surgeon who left the forceps inside you , the theatre sister who failed to count the instruments or the anaesthetist with the CO2 bottle who killed you, it is all the NHS's fault.

It may be that the only way of dealing with the automated car issue is to let the public drive them but not own them.

So a hospital buys a computer diagnostic tool from an IT company that says it diagnoses x. It makes a mistake, someone dies and family sue. The hospital definitely can't sue the IT company?

Why would we need any consultants or GPs or hospitals or consulting rooms to beat homeopathy on cost effectiveness?

there's literally no difference b/t robots and public sector workers anyhow

ok all jokes aside what work will be available for those people replaced by robots then...?

That's the problem we have to solve! Stop the robots ! I am sure the government will stop robots going past a certain point and if they don't we have a lot More problems than to worry about than jobs!