Which jobs will be last to be automated/outsourced?

This might be an IMHO threat, but I spend more time in GD.

The thread on banning automation got me thinking about which jobs would be last to be automated and/or outsourced to cheaper labor markets.

Here are my contributions:

Plumber – they have to deal with too many special cases in too many strange spaces for a robot to be successful

Electrician – see Plumber

Nurse – This is a hands on job that people will likely want to keep some human interaction, and can’t be done remotely (unlike doctor/surgeon – see below)
Unexpected jobs that may face significant reductions in force, outsourcing, etc., in my view:

Doctor – expert systems and AI may someday soon get to the point where they are better than humans. For those cases where a human is important, there’s no reason why a qualified doctor from, say, India, couldn’t provide the final OK. I think I read that Google has an AI that reads x-rays better than doctors or technicians.

Lawyer – Lots of contractual work may someday be done by an AI or expert system, or at least able to assist a lawyer, reducing the need for lawyers or paralegals. Accounting software has already done this for tax accountants, for example.

Surgeon – Robotic surgeons seem like a no-brainer to me, especially when assisted by a human nurse. For those types of surgery where this isn’t possible, it seems eminently feasible to outsource routine surgery to remote surgeons, getting help from robotics and nurses

Truck drivers, pilots, cab drivers, train engineers seem almost too obvious to discuss, when looking forward to the next few decades.

Thoughts?

(Note – I will be away from a computer starting soon for most of the rest of the weekend, so I may not be back to comment for a little while.)

There was an old Last Week Tonight episode about this where John Oliver quoted the definition work & automation experts opine will still be needed for the foreseeable future. Give me a second to look it up : there it is : “A series of non-routine tasks that require social intelligence, critical thinking and creative problem-solving” (though that should be and/or IMO - a plumber doesn’t really need social intelligence nor creative problem-solving for example)

A plumber absolutely does need creative problem solving. I suspect surgeons do, too, for many of the same reasons.

How about massage therapists, hair dressers, and prostitutes? Not a lot of creative problem solving or critical thinking, but I think customers put a high value on the human touch in all of those fields.

I guess it depends on what you mean by automated. All those jobs you mentioned already have some automation in them, after all, since they use things like expert systems, AI and other automation tools. If you mean so that only a robot does them, then I’d say something in the entertainment sector. Author, content creator, that sort of thing. Pretty much anything that has any sort of repetition leads itself to automation. Nurses could be automated, as not only could you use AI and expert systems to monitor people and administer medicine or treatments you could have robots do a lot of the chores nurses do as well. Same goes for doctors…they already are using automation to assist in surgery as well as AI and expert systems to help doctors track trends, and there is no reason that can’t or won’t continue.

That said, I seriously doubt we’ll ever get to the ‘last’ job to be automated, as I think what will end up happening is a further merging of human with AI/expert system and robotics. All have strengths and weaknesses, but combined I think what will be found is that they work better with a combination, not separately.

I suspect from a more white collar perspective, there’ll still be a fair demand for analysts who can basically interpret what the end users want into what the AI computer programmers will understand.

I figure if AI gets to the point where it can quiz a semi-Luddite guy working in the field for a public utility about what he needs the mobile device and computer system to do, and then translate that into actual system modifications, pretty much ANY job can be automated at that point. We’ll have AI that can use intuition, metaphor, analogy, etc…

Oddly I think everyday handymen will be a long time in being replaced, because a lot of the time, maintenance work like they do is basically applied creative problem solving. If it was all just measuring, cutting, screwing, etc… then it would be easily automated. But being able to identify what’s wrong and fix it on something that’s been repaired multiple times in the past with a mishmash of techniques and parts is something people can do, but machines currently cannot.

President.

https://news.sky.com/story/disney-unveils-donald-trump-robot-at-florida-park-11177652

I do think that dog groomer will hold out even longer than hair dresser.

A human would adapt to being worked on by a robot. I don’t know if a dog would be as easily trainable.

I suppose that depends on what we mean by “creative problem solving” - a body is a body is a body, the surgeon might have trouble cutting that one up right because some dumb vein’s in the way ; but he rarely has to invent an entirely new tool to cut it up or an entire new way to use a bloody knife. Same for the plumber, who might come across some weird configurations that require some kludging, or some ass backwards kludging that requires a complete re-configuration :slight_smile: ; but he rarely has to invent new techniques on the spot like, say, a software engineer, a particle physicist or a chef has to. Both plumber and surgeon take the majority their solutions to problems from a pre-learned, pre-trained toolbox, applying whichever one will/should fit their immediate circumstances.

That being said, yeah, I don’t see any construction job (architect, electrician, plumber, mason, locksmith, etc…) being replaced by bots any time soon. Unless we move to exclusive block construction and every single housing unit is the exact same, but I don’t see that happening at all.

Some folks pay good money for an extra critical prostitute. Or so I’ve heard. Not that there’s anything wrong with that.

That’s not a robot. That’s basically a big Chatty Cathy.

True, but I can see as technology progresses a bit, public functions could be done by tele-presence robot. A hundred years ago, If you wanted to see the president, you had to go to Washington, or he had to come to you. Now we have TV, and we can have the president visit out living room anytime we want. This is just the next logical step.

Save a ton on security and transportation costs, so the jobs lost to automation would be in protection and logistics services.

I think jobs will continue to change over time but the idea that automation will actually cause everything to be automated ignores human nature. In fact we’ve already automated thousands of times more capacity than existed when we started the industrial revolution. Production that only exists because of improved production tech that has been itself been made obsolete again and again over the last 200 years.

The current AI fad is also massively overblown. Yes software is getting better at certain fuzzy problems but we are a long way away from some kind of human replacement level AI. Up until that points these are simply more powerful tools for us to use.

In addition more jobs will appear because of these new tools. Jobs we can’t even anticipate. In 1992 nobody had ever even heard of a web designer for example, but it’s a highly paid job today that a lot of people do.

Auto mechanic and other technician jobs like hvac tech.

They require a high level of manual dexterity as well as cognitive skills. Robots lack the manual dexterity to do those jobs.

Also they pay better than minimum wage, but they don’t pay a huge amount so there is less incentive to buy a robot compared to a job like a surgeon or lawyer.

There is incentive to replace low wage workers because of the volume of work, and the ability of robots to work 24/7. There might be fewer techs, supervising dozens of robots. There will be (human) robot mechanics, whose jobs will be gradually automated until they themselves are largely superfluous.

Robots are dextrous enough to repair machinery and electronics now. They can fold clothes (but not socks yet) and pick delicate fruit without destroying the objects.

Here’s a recent thread from Hacker News where programmers and lawyers speculate on the utility of coding and automation in the legal profession:

Gigolo, and whore.
It’s quintessentially the human aspect that has economic value. Since youth is generally a professional prerequisite, there is a constant turn over, and the gradual automation of the pimp/madam aspects of the industry will have to stop at the low end of the hierarchy. Even exquisitely realistic human robots will fail on the cost/profitability equation, and elitist customer pretentiousness will ensure the market for a “real” human.

Tris


And the first shall be the last.

I’m guessing it’s a job that will be created by automation – one that doesn’t exist yet.

No matter how automated labor becomes, there will always be a niche for human service.

Even if a robot is better, there are some that prefer the pretentiousness of having a human do the work.

High end restaurants will still have human chefs, and they will still whip their meringue by hand, even if a robot can do it faster and better.

yes, you are right :slight_smile: .
But not just about those two professions: most jobs which require touching other people will not be automated.
A computer may be more accurate than a doctor at diagnosing symptoms,but the computer can’t stick a piece of wood down your throat to look at your tonsils.
A robot may be able to bring food to the table–but it can’t feed a baby, or an adult with dementia.
A robot cannot change a diaper.

So there will always be jobs for humans who take care of other humans.

Oh, I bet the robostitutes will have plenty of business.

Well, it certainly isn’t drafting, considering all I did to automate my jobs.

I predict that the rate of automation will actually slow down, because we have been picking the low-hanging fruit first. We will NEVER be at a point where even 50% of our citizens are unemployed due to automation. I’d be surprised if we saw the unemployment rate drop by more than a few percentage points because of automation.

I have been having this debate since the 1980’s in college. The robot apocalypse was always 10-20 years away. And 35 years later, it still is…

Humans have general intelligence and judgement. AI’s do not. They can do very specific things very well, but they can’t improvise or adapt in a general way.

There seems to be a misconception that workers are just drones mindlessly carrying out the plans set down from on high. That information flow in a company goes one way - from the planners and the deciders on top, down to the worker bees. If those bees could all be replaced by robots, we couldn’t need jobs any more.

But that’s not the way organizations actually work. Information flows two ways. Decisions get made on top, then when they are implemented there is a feedback loop where the workers either adapt to flaws in the plan (there are ALWAYS flaws in grand plans), or they tell their supervisors what can or can’t be done, and the plan is modified accordingly. Sometimes they actually find better ways to do things, because the planners don’t know what they know, and the process gets improved.

If you want to see what happens when the ‘worker bees’ just mindlessly follow the plan regardless of whether it makes sense, watch what happens to any organization when its workers go on a ‘work to rule’ strike. Even with a corporation, the day-to-day activities are far too difficult to perfectly plan, and managers absolutely require that people be able to modify things that don’t work or report them back. Robots won’t do that. Robots always ‘work to rule’.

I have visited many factories, and looked at many production plans. None of them actually match what’s going on on the factory floor. There are always little exceptions, things missed by the planners, that the workers are smart enough to work around. One of the hardest things about automation is covering the last 1% of exceptions that prevent humans from being taken out of the loop. Elon Musk tried to build a ‘dark’ factory that was just robots. He failed. He then admitted that automation is far harder than he thought it would be - especially for complex products.

Take the example of truck driver -a favourite candidate for a job that will go away soon. After all, when driving is automated, who needs human truckers?

Except truck drivers do a LOT more than drive the truck. They are responsible for making sure loads are loaded in order, secured properly, and correct. They act as agents for the trucking company, or are self-employed. They are security for the load, they learn to detect incipient problems with the truck. They deal with emergencies, and have to handle situations like road accidents where police are controlling traffic with visual indications. Sometimes they have to go off-plan, like when they get to a customer’s site and the load isn’t ready, or a vehicle is blocking the loading docks.

I could go on and on. A million little decisions requiring judgement. The ability to improvise when necessary. These types of characteristics are crucial in many jobs, even if 90% of the time the job is rote. It’s the times when it’s not that automation will fail you.

Also, you have to look at the environment around the truckers. Before you can automate you need fully digitized processes. You need common databases. You need commonality in loading systems, and probably new systems for allowing automated vehicles to handle things like construction zones, accident zones, road damage and the rest. An ad-hoc, patchwork system cannot be automated. But it can be run by people. A lot of our systems today are like that. Container shipping did that - they standardized processes to allow for automating of loading and unloading. Containers are all uniform in size, attach points, etc. It took decades to get to that level of standardization. And even now, lots of people work on those loading docks.

Will will see a LOT more automation. But we will also see demand for more workers. That’s been the pattern ever since the Luddites tried to shut down automated weaving, and there’s nothing fundamentally different about automation today - except that now the people being effected have social media to complain about it, and the jobs under threat now are the jobs held by influential people instead of, say, millions of agricultural workers.

Or as the saying goes, if immigrants were primarily made up of competition for politicians and lawyers, the borders would be closed tighter than a drum. Automation was good so long as it was destroying the jobs of farmers and assembly line workers, but now that it’s coming for the accountants, lawyers, professors and other intellectuals, suddenly it’s a major crisis.