Nurturing creativity with art, animals, and science fiction

Tag: Automation

To automate, or not to automate? The uncanny valley

A Glimpse of the Future

My mid-week post for the past two weeks has addressed a disruptive technology of major importance in the current global job situation: automation for greater productivity.

Robots and automated processes have already moved well beyond doing only what robotics expert Ryan Calo called “the three D’s: dangerous, dirty, and dull.” 

Last Wednesday I examined some of the ways robots and automation are replacing some types of traditionally minimum-wage or low-wage jobs, sometimes in appropriate ways, but other times in what some (including me) might consider needless, or less reasonable, ways.

Today I’d like to move up the social ladder a bit, because it’s not only blue-collar jobs that proponents of automation or robotics are proposing to pre-empt.

According to the research I’ve done for this series, doctors, nurseslawyers, financial advisorswriters, teachers, and child-care workers are also in the cross-hairs. At this rate, nobody can afford to get too smug. If professions requiring higher-level thinking and analysis are in danger from automation, NO job is safe.

Is that actually a real threat, though? Won’t there be at some point an “uncanny valley” effect? The uncanny valley is a problem in both animation and robotics. If you make something look or act extremely realistic–but just short of indistinguishable from the real thing–people react with revulsion. It strikes them as creepy.

The Uncanny Valley can be a scary place!

Could the uncanny valley save white-collar jobs? Well, maybe. The verdict is still out. There’s evidence that once people become accustomed to the almost-real look, they find it less repulsive. In other words, don’t count on it.

The end result SHOULD lie in whether the automation actually does a more satisfactory job than a competent human could. Meanwhile, this is a great source of thought experiments for science fiction writers, futurists, technological ethicists, and many others.

I’ve gotta say though, I find it interesting I haven’t yet seen any proposals that AIs should take over research chairs in the field of robotics research.

But think about it. Once we’ve reached the singularity, is there any career they’d find more interesting?

I’d bet not.

IMAGES: Many thanks to Before it’s News for the “looking to the future” graphic, and to G Financial Services Marketing for the “ranks of white-collar robots” illustration. I’m grateful to PandaStrike for the illustrated Uncanny Valley graph, and to HR Zone for the robot photo.

To automate, or not to automate? Is there value to the human element?

A Glimpse of the Future 

Last week I took a first look at some of the jobs that have been increasingly moving over to automation, and a few that might see more automation and fewer humans doing the work in the future.

In some cases this might not be a bad thing. In other cases, the robots may not do as good a job as humans might. A couple of cases-in-point leap to mind: bank tellers and retail store checkers. Which do you prefer?

Love ’em or hate ’em (I know people who feel both ways), these machines seem here to stay.

I’m older than dirt, so I remember before they had such contraptions. I remember having to plan to get money before the bank closed for the day or weekend, and how you always talked with a human being before you could complete any transaction.

I kind of liked it (confession: I still don’t own an ATM card, out of security concerns. Planning ahead: it’s a thing.), but then, I live in the Midwest, where bank tellers and grocery store checkers are apparently friendlier than they are in some other parts of the world. I like to get to know them, in the fond hope that if someone they didn’t know came in and tried to wipe out my bank account, they’d question it. I feel quite certain my bankers at Kansas City’s Country Club Bank would. Thanks, guys!!

I also remember before there was a self-checkout line at the grocery store. I even remember before they had bar codes on the groceries (what a pain that was!), and you had to watch the checker to make sure s/he didn’t make an error or ring something twice that you only bought one of. Of course, now when the machine steals your ATM or credit card information, you have few ways of knowing, so is that a net gain? Depends on for whom, I guess.

There’s reportedly now a trend toward automating fast-food service, unfortunately driven in part by the industry’s resistance to paying its employees a living wage. I can see how an automatic timer to pull the fries out of the hot oil at the penultimate moment might be a good thing, but completely removing all or most of the people? That’s a farther stretch for me.

You see, we’ve actually had automated fast-food delivery for a long time. They’re called vending machines, and they aren’t actually noted for their-high quality products or their ambiance.

Granted, Mickey D’s isn’t long on “ambience” either, but I kind of like to chit-chat with the smiling teens or senior citizens at the counter. Call me weird, but I prefer dealing with people, over figuring out the interface on yet another dang gadget. I’ve kinda perfected the human interface, at least to some extent, and I have this weird notion that people should be respected, even when they have low-end jobs.

An automated fast-food “restaurant” looks an awful lot like a glorified vending machine to me.

As I see it, the whole key should be playing to strengths. Robots and automation do some things way better than people. Business Insider interviewed Ryan Calo, a professor at University of Washington School of Law with expertise in robotics, who said, “For a long time, artificial intelligence has been better than us at highly structured, bounded tasks.” All of the applications we’ve looked at so far in both this and the previous post on this topic have been in that category.

Calo thinks, however, that robots are now, or soon will be, capable of moving beyond “the three D’s: dangerous, dirty, and dull.” It’s a fine line to define (sorry for the rhyme), so where do we draw it? If robots and automation can lift us beyond those “dangerous, dirty, and dull tasks,” isn’t that a net gain? I think it definitely is. If they can ever design a Roomba that cleans the potty, I’m all in!

Ivan Fourie encountered this friendly store clerk in Kyoto 2006, and immortalized her in a photo.

But people right now (and for millennia) do/have done way better at some things than robots and automation have managed so far. The determination to push automation/artificial intelligence beyond those basic limits won’t stop. (we’re talking about humans with an intellectual challenge. Of course they’ll pursue it as far as they can).

But just as industry doesn’t want to talk about the full cost of their initiatives (including environmental and human damage), so the people involved in the “second machine age” don’t want to talk about ALL the costs of their initiatives.

Are these Chinese robots cute enough to be worth their cost in human devaluation? Are they worth the effort of putting “friendly store clerk” and her siblings all over the world into financial devastation?

Would their AIs put good people out of work that they need? Don’t we all need people who are a positive part of their community? The friendly 7-Eleven clerk who brightens our morning? The bank teller who keeps our accounts safe? The shopkeeper who grows her small business locally? The first-generation immigrant family who runs the gas station? The custodian who keeps the school clean and well-maintained?

What’s the human cost of the fancy machines? Do they make life better for the humans in the community, or only for the corporations running the businesses?

I think we’re at a crossroads, in our contemporary life. We can look globally at ALL the costs of the decisions we take, or we can keep on looking only at money in a system skewed to ignore some of the most important costs of all.

Our choice.

Our future.

IMAGES: Many thanks to Before it’s News for the future-vision graphic. The photo of the Safeway self-checkout is courtesy of WonderHowTo, and the photo of the ATM machine is from The Northeast Today; many thanks to both of you! The cynical minimum wage meme is from Ron Paul’s “Liberty Report.” Your thanks is that I acknowledged where it came from, dude. You certainly illustrated my point, anyway. Many thanks to NPR’s “All Tech Considered” for the photo of the automated fast-food restaurant. I am grateful to Ivan Fourie’s Flickr Photostream for the the friendly store clerk’s photo. Many thanks to Business Insider for the photo of the Chinese food service robots.

To automate, or not to automate? First thoughts

A Glimpse of the Future

Automation and robotics have been making a lot of inroads, lately, and the trend seems unlikely to change.

Unless you live in a particular fringe political neighborhood, you’re aware that the vast majority of those jobs have been taken by automation designed to boost productivity. According to one study conducted by Ball State University that looked at manufacturing job losses in the US between the years 2000 and 2010, 87% of those jobs went to productivity, but only 13% to trade.

Losing one’s job to a robot is far more likely than losing it to workers in Mexico or China. It’s not really a new story. As the Ball State study illustrates, it’s been going on for decades, and not only in manufacturing. An NBC article from several years ago gave a list of jobs that humans could lose to robots in the near future.

The trend is spreading, inevitably. Sometimes it’s a good thing. Astronauts might be sexy, but humans shouldn’t be the first things we send to new places in space. I think I speak for many when I say I’m happy to see Spirit and Opportunity go to Mars before people brave its hazards.

A NASA artist’s conception of the Mars Rover Spirit.

An argument could be made that automation in certain sectors makes things faster, more efficient, and less error-prone. My husband works in a diagnostic lab where processors and stainers perform many routine tests that once were run by hand. An argument could be made that pharmacy automation might be less subject to corruption or error (though there are many ways humans could take advantage if the system isn’t carefully set up and monitored).

This is an automated slide-maker and stainer for a specific laboratory purpose. It delivers consistent results that would be hard to achieve at speed by hand.

But here we start to run into a gray area. There still are things it seems likely robots are a long way from being able to do as well as a skilled, trained human.

My husband, for instance, is still the experienced tech the doctors call upon to quality-check for diagnostic results whenever they do certain kinds of biopsies.

Automated pharmacy equipment from RoboPharma (yes, that’s really their name).

The pharmacy robot may be able to package up pills at the speed of light, but how will it do when you need advice about the best cream to use for the persistent itch you have, or which syrup might work best for your baby’s cough?

I plan to explore this question in more detail next week.

IMAGES: Many thanks to Before It’s News for the “future” graphic, the CNN Money for the graph illustrating the Ball Stat University study, and to Electronicsb2b for the photo of the robotic automotive assembly line. Many thanks to the invaluable Wikipedia for the image of the Mars rover Spirit, to Abbot Labs for the photo of their Cell-Dyn SMS slidemaker and stainer and to RoboPharma for the photo of the automated pharmacy equipment.

Powered by WordPress & Theme by Anders Norén