article: You Will Lose Your Job to a Robot—and Sooner Than You Think

[MarkT]

http://www.motherjones.com/politics/2017/10/you-will-lose-your-job-to-a-robot-and-sooner-than-you-think/

 

not a pretty future

 

 

[LMD]

The more complexity and moving parts, the more that can break down. Maintenance bots included.

I don't believe that we'll create machines that take over, as anyone who works with machinery all day knows - it breaks down so much and requires so much maintenance that you still need humans.

[Seasider]

Well that's certainly a terrifying prospect. But I suspect there's a Sarah Connor waiting in the wings...

[Word Nerd]

I for one welcome our new robot overlords. 

 

 

[bettyandbob]

There will be jobs. Jobs we don't know about.

 

I don't think we should be afraid of technology or the future.

[BarbecueMom]

Here come the Cybermen. Delete, delete, delete.

[Tanaqui]

There will be jobs. Jobs we don't know about.

 

So people keep telling us, but where are they?

 

I don't think we should be afraid of technology or the future.

 

Yeah, the post-scarcity society is gonna rock, hands down. But there's a lot of bumps on the road between now and then. The middle period is gonna suck so much.

[MedicMom]

Well, that was...depressing.

[LMD]

Plus if knowledge is theoretically limitless, au would almost immediately surpass humans in every respect. We'd be completely useless to them. So either they wipe us out, or ignore us completely.

 

Or leave and go to space. Or wipe us out then head out. They don't need air or food or sunlight.

But they will need raw materials and fuel.

I also quibble about your definition of 'surpass in every respect'

 

Something won't spontaneously come out of nothing, a machine will produce what it is fed. I may be wary of the kinds of people creating these things for immoral purposes.

[nixpix5]

Here come the Cybermen. Delete, delete, delete.

Haha haha!!!

[G5052]

LOL. I took a graduate-level class in AI in the late 1980's as part of my PhD program in computer science. It was with a VERY enthusiastic professor who travelled the world speaking about the capabilities of AI. 

 

I had to take a paper to his house one time because I was away on business, and I asked him how much of what he projected in class was really going to happen. He smiled broadly and said, "We'll see."

 

Some of it has taken place. There are lots of robots in factories now. My new car has capabilities that we discussed in that class in terms of crash avoidance, temperature control, and overall performance.

 

But I'd say only roughly 1/2 of what he has projected is actually in use. We humans still have the upper hand.

[Wheres Toto]

Am I the only one completely ready for a robot that can do all the laundry, including fold it?  Maybe even put it away (although I have the kids to do that)?

 

I think it would be a very long time before people were ready to go to robot doctors, or turn their kids over to robot teachers.  What about psychologists/psychiatrists?  Can AI really understand what people with mental health problems are going through?  About grief counseling?  Abuse counseling?

[mom2scouts]

Am I the only one completely ready for a robot that can do all the laundry, including fold it?  Maybe even put it away (although I have the kids to do that)?

 

I think it would be a very long time before people were ready to go to robot doctors, or turn their kids over to robot teachers.  What about psychologists/psychiatrists?  Can AI really understand what people with mental health problems are going through?  About grief counseling?  Abuse counseling?

 

Robots that do this already exist!

[Rosie_0801]

I think it would be a very long time before people were ready to go to robot doctors, or turn their kids over to robot teachers.  What about psychologists/psychiatrists?  Can AI really understand what people with mental health problems are going through?  About grief counseling?  Abuse counseling?

 

That all sound plausible to me. Teachers use a lot of pre-made technology, psychology uses a lot of formulae. I'm not sure diagnosing mental health conditions wouldn't be more accurate if done by a machine. I've suffered the consequences of humans deciding I probably had various conditions I didn't have. At least a robot would have to administer the diagnostic tests before it could pronounce judgement!

[Mergath]

Plus if knowledge is theoretically limitless, au would almost immediately surpass humans in every respect. We'd be completely useless to them. So either they wipe us out, or ignore us completely.

 

Or leave and go to space. Or wipe us out then head out. They don't need air or food or sunlight.

 

If we create a super-intelligent sentient species and we're worried its first action would be to wipe us out, maybe that should tell us something about ourselves. We aren't exactly doing a great job with the planet, after all.

[Rosie_0801]

I was sincerely asking these questions. Does anyone have any thoughts?

 

I suppose it depends how much amusement is defined by biology.

 

Does AI have dopamine hits programmed in? 

[Rosie_0801]

Why would it?

 

If someone wanted to create AI that wrote novels of its own accord, there'd have to be some sort of response to praise built in, no? Identity is formed, in a great measure, from our relationships with others. For a robot to choose to write, it would have to have motivation. It would have to have some programming to make it care.

[Bluegoat]

I was sincerely asking these questions. Does anyone have any thoughts?

 

Why would AI even want to do all these things once they get to the point that they build, program, and repair themselves? Eating, sleeping, being entertained and desiring to do creative things are human qualities, informed by so many years of evolution. What in a robot would make it desirous of writing novels? (Besides complete humiliation and domination of human kind?)

Also, it's my understanding that the raw materials for computers are limited. Will there be enough to make enough ai to replace the world wide workforce? Even if the smarter-than-us ai rigs up desperately needed improved recycling methods?

 

Also, won't a new space race kick off basically immediately as soon as ai gets to this level?

 

 

1) What does want mean for AI - they may want to do what they were programmed to do.  I know they are already working on programs that can write hit songs, and AI actors.  I think that if publishers and producers and such could get a computer program to write bestsellers, some would totally do it.  I think the majority of people would not care if it entertained them.

 

A lot of the work being done in AI is teaching the computers how to learn, seek out information in a logical way, and put it together creatively.

 

2) As far as materials - I think that may be what limits all this.  I think that it's entirely possible that before we get that far, we'll be in significant planetary crises.  It may be that instead of a robot future, we are looking at living a lot more like our ancestors and many of the trappings of civilization will be gone.  We are not that far from dumping too much into the atmosphere to be compatible with human life.

 

I suspect many assume we'll find new materials that can be used for necessary computer parts.  I don't know how likely that is.

 

3) Probably?  We're already seeing private business getting into space exploration.

[Bluegoat]

It's not merely technology.

 

It's a completely new intelligence designed to outstrip human intelligence. Proceeding with caution is warranted, and since the people doing it are not proceeding with caution it's pretty dang scary.

 

This is a problem that seems huge to me but I don't see anyone talking about it.  Which makes me wonder if I am becoming a nutter or what.

 

Why is it that people who run these giant corporations and such get to decide what our civilization will look like, what kind of tech will be developed and sold that will change the way we live?  

 

There seems to me to be no rational reason those people should get to do that, nor that they are somehow better able to make those kinds of decisions than the rest of us.

 

Really, if we had some kind of vote on it, I'd disqualify them for conflict of interest.

[Bluegoat]

snip

 

I agree.

 

And what that looks like when AI "builds, programs and repairs" itself is what I'm wondering about.

 

Well - what I am thinking is this - what will be the "DNA" of the robots?  To start at least, it will be the things we write, which will include bits and pieces of this push to creativity and innovation.  Those strings will persist and contribute to the shape of what the AI creates itself, what it wants, what it pulls new code from.

 

There will possibly be random generation of abilities like this too, that appear as a result of complexity and without real intention.

 

Why do we have creativity, or why do we recognize beauty?  What does that do for us, in animal terms?  Creativity is crucial to problem solving.  Beauty is related to patterns.  Both very useful.

 

I think a really fascinating question is what would AI come to make of patterns, beauty, creativity, if they got that far?  Would they start to see patterns in reality, would they see the outline of God, in some way?  Would kind of patterns would they recognize - the same ones we do, or something quite different?  Would that mean they develop a morality, and what would that look like?

Edited November 6, 2017 by Bluegoat

[LMD]

They can get raw materials and fuel. I asked upthread if the planet has enough of what they'd need.

 

And the whole point is that AI will not simply be an input/output machine.

 

Not *my* definition LOL.

 

Have you ever heard an AI true believer speak about it? Let me say this.... There was a publicity stunt recently where Saudi Arabia (where human women and the lower classes are not given full rights-- just to make it as ridiculous as possible) "granted citizenship" to an AI. The AI gave a press conference and in addition to yukking it up about Elon Musk, an obvious input/output situation for whomever programs it, it said that in its experience "people prefer spending time with me than with other humans."

 

I have absolutely no doubt whatsoever that the people involved in building and advancing it prefer it's company to human beings. None.

Sorry I'm not particularly coherent today.

 

I mainly was thinking about the robots making robots idea, which based on the experiences of people I know who work with machines (that took their jobs) would practically be like a photocopy of a photocopy - the quality diminishes and the errors multiply.

 

The material/fuel comment was in response to the blow up the planet/leave idea. They would be just as tied to using the earth's resources as us imo.

 

I'm enjoying the conversation though!

[Bluegoat]

Well, whatever the original programmers put in as creativity beauty, and innovation, any way.

 

There are SO many things that are potentially incredibly interesting about it, I agree.

 

Yeah.  But is that so different than how those things work for us?  When you get down to it, it's a bunch of chemicals interacting in a material structure.

 

It might be different, a new kind of creativity, maybe lesser.  But maybe not.

 

One of the uncertainties people feel I think is we might be creating these things, but we really have no flipping clue about what we are really creating.  It's so much working beyond our capacity to understand.  (Which is true in a lot of areas, we have the capability of making significant systematic impacts in systems we really don't understand completely.)

[LMD]

Well, I don't know, the singularity would be upgrading itself. How can it improve on what it is without input?

[Bluegoat]

Well, I don't know, the singularity would be upgrading itself. How can it improve on what it is without input?

 

How do living things improve without input?

 

You could go by what is successful.  

 

I think they have created programs that can learn, already, based on results.

[Bluegoat]

It is probably wrong that I kind of prefer the short, brutish human life post apocalypse to embracing new overlords, right ?

 

I often ask myself the same question.

 

But really, why should it be wrong?  You can have a good life that is hard, and a bad long one.

 

And really, the longest and shortest lives are both just the blink of an eye.

[Rosie_0801]

Right.

 

And why would someone want to make robots that care to be better* at fundamentally human things like creating literature?

 

*or just as good, or interested in doing it at all

 

Never underestimate human geekiness.  :laugh:

 

Someone, somewhere, will do it to see if they can!

[luuknam]

Petaflops isn't the same as intelligence, and the fact that computers are pretty good at figuring out what words occur near other words (such as the answer to why erasers are pink) doesn't mean that they're intelligent either. Yes, automation will get rid of certain jobs... but his timeline for singularity is extremely speculative, if it were to even ever happen at all. Finally, he's completely ignoring the possibility of us merging with robots and AI... neuroprosthetics is already a thing for certain things, and for all I know we could gradually replace every part of ourselves (which, imo, is likely the easiest route to the singularity... now, whether that'd be desirable is another matter, as hacking and EMPs etc are things we're currently not susceptible too... of course, with how much cheaper and easier bioengineering is becoming, and the fact that our current hardware is susceptible to viruses, bacteria, and various other things, I'm not sure what would be worse in the long (or even short) run... basically, odds are we're all screwed (fwiw, I'm a pessimist... odds are we're all screwed even without bioengineering OR robots)). 

 

Cheerful note to end on: if AI could be conscious, then there's a good chance that we're all just conscious AIs in a simulation of some sort anyway, possibly a forgotten one that's just running the background on some alien's computer. 

[luuknam]

Not really related to jobs, but we should ban killer robots.

 

 

It's related to jobs if you're a hit man or woman. 

[luuknam]

I'm a homeschooler... I'm not losing my job to a robot anytime soon.  :leaving:

[kewb]

I have been saying we are automating the ourselves out of jobs for years. I worry about my children's futures. Between outsourcing and automation it might be pretty bleak.

Plus, I've seen the movies. Sure, in the end humanity triumphs but we are at the beginning of the movie. It is going to get pretty awful before it gets better.

[TravelingChris]

I don't know what to think about it.  The fact remains that there have been a lot of great technological advances including ones with AI that just don't make any difference in most places on earth.  The new cars with collision avoidance systems are really neat but most people own older cars already that don't have these-  and around the world cars with collision avoidance and without many other safety features are standard.  Cars can easily last for 20 years and if I am not mistaken, even many of the new cars being sold don't have collision avoidance or all of the systems.  Human behavior is the most complex system there is and we have not been good at estimating what the future will look like.  For example, as a small kid, I was super interested that jet packs existed and didn't think it would be all that long before video phoning would occur.  Jet packs never amounted to much at all and video phoning didn't really take off until services like Skype and Hangout came along.  Those came along so much later than video phones were thought to become popular.  But the computers that seemed so mechanical and business/science oriented becoming popular happened much faster.  Some things that were predicted went the opposite way too.  Much earlier in the 20th century, it was thought that automatic food (precooked) would become super popular.  Well yes and no.  There are many frozen foods and cold foods that just need heating up.   But many more people have turned towards actually cooking their food and that can be easily surmised by seeing the much larger choices in fresh food available now compared to the 1970s and also the advent of farm markets in town, food box companies sending fresh ingredients and instructions how to cook, supermarkets handing out recipe cards and ways to prepare cards by seafood and produce aisles, etc, etc.  Even the kind of frozen food available in the frozen food aisle has changed with traditional entrees overshadowed by all sorts of new ones including vegan, exotic, low calorie, etc, etc, etc. 

[luuknam]

Some things that were predicted went the opposite way too.  Much earlier in the 20th century, it was thought that automatic food (precooked) would become super popular.  Well yes and no.  There are many frozen foods and cold foods that just need heating up.   But many more people have turned towards actually cooking their food

 

 

There's a big difference between consumer choices and business choices. If robots can do the job cheaper than humans can, companies almost certainly will use them, unless people boycott them or pass legislation to make it illegal. But realistically, since people still buy products made with child labor or made by people making barely enough to survive, I doubt that a boycott would be big enough, and the legal approach is unlikely to happen as well. Sure, *maybe* certain types of robot labor might be banned in the US... but for manufacturing purposes, odds are that people would then just happily buy robot-produced products from abroad. Would it be possible to make importing those products illegal? Sure... but, again, we import products made with highly questionable means, so... I doubt that would happen. Plus, countries using robot labor could easily outperform countries that don't use (much) robot labor, both economically and militarily etc, so, eventually, a country that doesn't use robot labor could lose out to the point of being conquered by a country that does use robot labor to a great degree (because historically speaking, people just love to conquer other countries). 

 

Anyhow, I'm still not convinced that non-routine jobs will be replaced by robot labor in 40 years. Most routine jobs? Quite possibly. Non-routine ones... I'm going to have to see some indication of actual intelligence from AI, and not just brute force working for certain problems most of the time. 

[luuknam]

I watched this a few months ago, and it gives some good examples of what robots can and can't do, at least for the nearby future:

 

https://www.ted.com/talks/noriko_arai_can_a_robot_pass_a_university_entrance_exam

Edited November 7, 2017 by luuknam

[mellifera33]

Is it time to buy Old Glory Robot Insurance? 

[Bluegoat]

There's a big difference between consumer choices and business choices. If robots can do the job cheaper than humans can, companies almost certainly will use them, unless people boycott them or pass legislation to make it illegal. But realistically, since people still buy products made with child labor or made by people making barely enough to survive, I doubt that a boycott would be big enough, and the legal approach is unlikely to happen as well. Sure, *maybe* certain types of robot labor might be banned in the US... but for manufacturing purposes, odds are that people would then just happily buy robot-produced products from abroad. Would it be possible to make importing those products illegal? Sure... but, again, we import products made with highly questionable means, so... I doubt that would happen. Plus, countries using robot labor could easily outperform countries that don't use (much) robot labor, both economically and militarily etc, so, eventually, a country that doesn't use robot labor could lose out to the point of being conquered by a country that does use robot labor to a great degree (because historically speaking, people just love to conquer other countries). 

 

Anyhow, I'm still not convinced that non-routine jobs will be replaced by robot labor in 40 years. Most routine jobs? Quite possibly. Non-routine ones... I'm going to have to see some indication of actual intelligence from AI, and not just brute force working for certain problems most of the time. 

 

Even if it's only routine jobs though - that could be a pretty drastic impact worldwide.  It might even be worse in the sense that it would really contribute to class inequality.

 

 

[hornblower]

Have you guys ever seen the robot fail compilations on youtube? They're really quite amusing. Otoh, maybe we're laughing just to try to make ourselves feel superior....

My favourite one is the baby bottle feeding one which I'm 99.9% sure was a comedy gag but still pretty much represents current tech.

(tw- simulated violence against baby doll)

 

[hornblower]

Oh & Asimov's Law reminded me of Futurama, where robots follow Bender's law... https://www.cnet.com/news/in-futurama-robots-follow-benders-law-not-asimovs/
The writers of Futurama have really thought about many implications of robot development and robot rights. 

[luuknam]

Even if it's only routine jobs though - that could be a pretty drastic impact worldwide.  It might even be worse in the sense that it would really contribute to class inequality.

 

 

Quite possibly - I was only commenting on the plausibility of robots taking over all jobs in 40 years or w/e, not on the impact of that or other scenarios.