Fresh Air with Terry Gross on NPR did a segment this week about this topic: “How Close Are We Really To A Robot-run Society?”
Educators in Australia are starting to see the implications of automationfor their students and to sound a warning. Fifty percent of Australian students, they say, are trai9ning for jobs that either will not exist or will be radically transformed.
Thanks, Tangent–I will check that out.
Interesting, Captor. I know I have been careful to try to steer my kids away from training for careers that are likely to be either extinct or made so efficient as to reduce the demand dramatically. And one of those, IMO, is an area that so many people (including policymakers) think is on the opposite side as an area of future demand: computer programming (being a “coder”). I strongly suspect in twenty years max, an AI will do all the actual coding, and creating software or “apps” will involve just talking to the UI about what you want it to do. That may still require some jobs for creative people, but all the learning of programming languages and so on will likely be a waste.
An interesting link. The interviewee was interesting in that he “got” that self-driving cars are a lot safer than the other kind, but I don’t think he understands that the way a computer “learns” something can be applied to all other computers in its network instantly will make them so MUCH better than humans, even in emergencies. Very rapidly, cars driven by humans will be regarded as dangerous and unpredictable death machines.
Speaking of NPR and self-driving cars, Planet Money did a good report on this. It seems that Google faced a choice: emulate airplane autopilots (where all the controls are still there for a manual takeover), or elevators (which initially scared people when no operator was in them)? They initially tried the airplane style, but found this was actually much more dangerous. So now they are trying to get California to let them remove the steering wheel.
This is Google Glass hubris. Elevator riders don’t buy elevators, so their opinion about control doesn’t count. Also, elevators have an easy failsafe option - cars, not so much.
Markoff has been covering Silicon Valley for the Times for a very long time - don’t underestimate what he gets.
The last thing we need is for this type of networking. First, thousands of cars communicating to millions of cars instantly is going to be a bandwidth nightmare. Second, the new advice has to be integrated with the old advice in a coherent and consistent way. Third, if you let a car transmit “if you see a kid, stop” you can’t prevent a hacker from transmitting “if you see a kid, speed up.”
I’ve had the firmware in my Prius updated a few times. It is done at the Toyota dealer, not while I’m on the road. That is a good thing. You really don’t want to reboot in the middle of a freeway.
Here’s an interesting ranking of pretty much every job you could think of (and many you might not) from least to most likely to be automated by the year 2033.
I found it odd that most hands on blue collar jobs had an 80% or better chance of being computerized but tech school teachers only had a 26% chance of being computerized. Who will be taking classes for jobs that don’t exist?
That’s an interesting point. But what you are talking about could be called indirect or second-hand computerization. I think they are limiting themselves to cases where a computer will actually be doing the job, rather than getting into ripple effects. It’s a fair point though and at least one degree of “ripple” should be accounted for in cases like this where it is so clear.
Can anyone explain why kindergarten teachers are so much more “computerizable” than non-special education elementary teachers? And why aren’t special ed elementary teachers listed? My wife wants to know how secure her position is!
I’m having a lot of trouble wrapping my head around kindergarten and elementary school teachers being automatable at all. It seems to me that both sets of jobs demand extremely good social skills, which computers are likely to lag far behind on, simply because they are computers and not human.
Some cars are already hackable. Clearly this is an issue we will have to deal with, and soon, driverless cars or no.
I agree, and these researchers seem to pretty much agree when it comes to non-sped elementary teachers: they give them a likelihood of less than one-half of one percent of being computerized. With kindergarten teachers, they still see it as significantly more likely they won’t be than they will be, but the chance zooms all the way up to 15%, 34 times greater than for other elementary teachers! 
If you get into secondary effects, then who knows? It does seem ridiculous that regular teachers can’t be automated, but kindergarten teachers or preschool teachers could be.
However, automation doesn’t always work the way we think it does. We don’t have robots that change diapers, but disposable diapers made in a factory and sold in a grocery store reduce a lot of the work it takes to clean a baby’s ass.
Turn on a TV set and let the kindergartners watch TV all day and hey, you’ve reduced the need for kindergarten teachers. Or maybe with everyone out of work since they lost their job to a robot, there won’t be any need for child care workers since unemployed parents will be at home and can look after their brats themselves.
I would have reasoned the opposite way! Very little children require emotional contact, nurturing, support, empathy, and all the “ape to ape” interactions that people excel at but which machines can’t (yet) be programmed for.
Yet older kids can settle down into “game” interface learning programs just fine, and so can be taught by AI even at the level that exists today.
Here is how all their teachers rank, in order from hardest to computerize to easiest (and if you move the decimal over two spaces in the number after the ranking, you get a percentage likelihood):
I can’t make rhyme or reason out of this. Why wouldn’t kindergarten be right up near preschool and elementary? And why is middle school fairly high risk, while both elementary and secondary are low?
Yet the list overall seems like it makes sense: up above 90% are jobs like Insurance Sales Agents and Conveyor Operators and Tenders, while under 10% is stuff like Sociologists and Fashion Designers; right around 50% is careers like Dental Assistants and Embalmers. All of those seem to fit. So what is up with these teacher categories being all over the place? And why aren’t elementary special education teachers listed?
That was a bug. When you make on-line access to the computers that drive your car a feature, then we are in for real trouble.
If you think the natural tendency isn’t for cars to become more connected, driverless or not, you got another think coming.