Japan is not yet the land of robots

Japanese hotels and banks are, by global standards, heavily overstaffed despite the country’s demographic crunch. Most supermarkets have not embraced the automated checkouts common elsewhere, nor airlines self-service check-ins. The offices of Japan’s small and medium-sized enterprises are among the most inefficient in the developed world, chides McKinsey, a management consultancy.

Japan has an elaborate service culture, which machines struggle to replicate. Japanese customers, especially the elderly, strongly prefer people to machines, says Yoko Takeda of Mitsubishi Research Institute, a think-tank. Employment practices make it difficult to replace workers. And while gimmicky robots abound, Japan struggles to develop the software and artificial intelligence needed to enable them to perform useful tasks, says a report by the Ministry of Economy, Trade and Industry (METI), the cockpit of Japan’s post-war miracle. So while the reception at the robot hotel is automated, seven human employees lurk out of sight to watch over customers and avoid glitches. Robots still cannot make beds, cook breakfast or deal with a drunken guest who will not pay his bill.

Here is more from The Economist.

Comments

The clever Japanese are working on it however.

Bonus trivia: GM 'robotic-sized' its factories in the 1980s, and lost a lot of money, because they were too 'leading edge'. The Japanese took their time and came out ahead when the technology matured more. Source: personal observations and from what I've read.

Such inefficiency of Japanese domestic sector was mentioned in Raghuram Rajan's Fault Lines. One particular example he shared was that even today, one can find people in Japanese corporations and hotels whose only job is to direct people to the nearest lift which has arrived on the floor.

That's not a bug, it's a feature, in five star hotels. With a declining workforce, Japan could not afford to do that in a one-star hotel.

Japanese people and some others who live here like having a biological person to talk to when there's a problem, instead being jerked around by a computer whose program doesn't address the specific issue, as for example with Bank of America's online banking site (just an example). Here, we walk over to the bank and someone fixes the problem.

if you're a customer with large deposits, fine. otherwise it's a waste of resources.

Ie, if you don't have a million bucks, a robot is much better at taking all your money and making sure you never get any back.

Doesn't matter how much money you have. Japanese banks care a lot about their local reputations. In any event, Japanese people and some foreigners here want to be able to address their complaint or question to a biological entity, not a program. A Japanese bank won't lock you out of the online system and then classify your account as abandoned (because you didn't access the system that you were locked out of despite repeated attempts by international telephone to get back in) and then send your money to the state government and then make it impossible to address the problem other than by a program that in fact doesn't address it.
Obviously this is the way the world is now, just saying Japan isn't fully on board.
Mulp's assessment seems pretty spot-on.

Now I'm truly curious. In Japan, how much does it cost per year to have a savings account + credit card + internet banking? I pay ~$80 a year. During the last 5 years I've talked to bank person 3 times, I think that's why the bank can offer such low prices.

As an expat not always proficient in the local language, I prefer the internet bank and support over the phone. There's not point int talking to a human that doesn't speak what I speak. Over the phone I may be lucky and get help in some language I know, walking into a bank branch may be even more frustrating than a phone call.

It costs nothing for savings, credit card, online assuming you meet the requirement to open an account. If you have language issues, that's that, the point is there is a person you can talk to rather than a program that will jerk you around and waste your money on international calls (trying to fix the problem) and not solve anything (if the problem isn't programmed in). If you live in Japan, go to the bank personally.

Well, Japan is a land where they can build robots at least, in contrast to some other places.

'What happens when we stop the flow of knowledge up the stack? I think that the weakness of the US industrial robotics sector is instructive. The US has little position in making high-end precision manufacturing equipment. When it comes to factory automation systems, machine tools, robot arms, and other types of production machinery, the most advanced suppliers are in Japan, Germany, and Switzerland.' https://danwang.co/how-technology-grows/

Well, Australia plays a role in robotics development despite having an industrial base that is mostly hollow. That is, we only tend to make hollow things here on account of how their bulky size makes them expensive to ship in from overseas. So the United States will have a part to play. It would be hard not to given its size.

Well, here is the next sentence from that citation - 'I think the reason that the US has little position can be tied directly to the departure of firms from so many segments of manufacturing. How do engineers work on the design of automation systems if they don’t have exposure to industrial processes?'

Because in the 1980s, the U.S. was certainly a leading nation when it came to automation/robotics - 'In the 1980's, automotive companies showered robotic companies with investments. The enthusiasm and funding were not always matched with understanding. General Motors Corporation spent more than $40 billion on new technology in the 1980's, but a lack of understanding led to costly robot fiascos. In 1988, robots at the Hamtramck Michigan plant wreaked havoc - smashing windows and painting one another. Unfortunately, the premature introduction of robotics began to create financial instability.

It wasn't until recently that the robotics industry has regained mid-1980 revenue levels. The American robotics market disappeared as Japanese and European bought up companies.' https://www.robots.com/articles/industrial-robot-history

The U.S. played a part a generation ago - these days, its main part is to buy products from other countries.

And let us be brutally honest - Japanese, Italian, Korean, or German automakers with manufacturing operations in the U.S. will continue to support those companies able to make robots. And for the Japanese and German automakers, the basic bet to make is that they will use automation suppliers which they already have a working relationship with - in Japan and Germany.

"the most advanced suppliers are in Japan, Germany, and Switzerland."

2/3 of these places with overreaching labor laws.

And 2/3 of those places just happen to headquarter the world's largest and second largest automobile manufacturers.

And I would have thought that 3 out of 3 of those places had overreaching labor laws, certainly by American standards.

Nothing surprising here. Human interactive robots are hard to do. That's why when you watch a movie about robots, apart from a few mock ups for close up shots they are all computer generated images because it's far easier than using actual robots. They explaining that to someone from the 1950s.

And shouldn't we be looking at Sweden, Germany, and Australia with their much higher minimum wage for non-gimmicky software and robo labor saving in service industries, rather than Japan?

A wall display that guides you to your room in a hotel ain't a robot but the 5 cents in wages it saves per guest per day adds up. It adds up to $18.25 a year. ($18.30 in leap years.) Throw in voice activation in each room and people will be too busy trying to get it to understand them to bother the staff.

If you're only tipping the bellhop five cents, I think you're a bit of an outlier.

In Australia there are no bellhops and no tipping. It's paradise.

Or at least it is if you travel light.

"Robots still cannot make beds" they are closing to it: the problem, that to predict movement of soft objects the computational power is currently too high, but still computational capabilities are growing.

cooking breakfast - it is also coming https://www.youtube.com/watch?v=M8r0gmQXm1Y - it is just not very developed now, but not difficult to see that this can be improved.
as for drunken guests - no there is no need for robot kicking humans - just institutional rules - allow to withhold money from account of drunken guest by providing some proof ( here those institutional economists who otherwise spend their time telling fairy tails could eventually make some intellectual effort and propose least harmful way to organize the process )

That video clip was painful to watch. But those exaggerated movements will soon be a thing of the past. I think Australia's greatest contribution to the field could be robotic laziness that results in them wasting less energy with their movements. (I would look it up to see if that's actually true, but my culture forbids me from making the effort.)

"they are closing to it: the problem"

I really doubt that they are close to replacing maids -- making beds, straightening up, wiping surfaces, hanging towels, noticing damage, retrieving lost items, etc, etc. Robots aren't close to doing any of these things, let alone all of them. Why do we persist in so undervaluing human flexibility, dexterity and hand-eye coordination? And the robotic egg cooker is hilarious. An expensive machine to do just one stationary cooking task very slowly? No, it is not easy for me to see how this can be readily improved to be practical any more than this did:

http://www.shorpy.com/node/23997#comments

The past is full of failed futurism. Robots in factories are one thing -- where a single, well-defined, stationary task can be repeated thousands of times. Robots doing many different tasks in unconstrained environments like homes and hotels is completely different (and, no, we're not close).

These aren't the droids you're looking for.

"METI ... the cockpit of Japan’s post-war miracle" Long ago, when I read the Economist regularly, it often paid homage to METI. I've always wondered whether it over-rated its role.

"Japan struggles to develop the software and artificial intelligence needed ...": what does that matter? Buy it in if you need it.

The Economist bought into the METI hype story, which was popularized by a political scientist, Chalmers Johnson, who didn't understand much about economics. Then again, one of their editors, Bill Emmot, went against the grain of popular thinking when he wrote "The Sun Also Sets" in 1991, soon after Japan's stock market bubble burst.

Keep in mind that the journalists who write on Japan for the Economist usually don't have an economics background either.

Another oddity about Japan is that it is common to regard urban residences as if they were disposable. The propensity to raze residential structures and build anew is a great deal higher than it is elsewhere.

It would be agreeable for all concerned if the peoples of the Far East had more children. No clue how the current state of affairs came about.

I take it you've been to Japan and seen the mud walls they still use to make houses, very green (or brown). The latest effort is more modern: super-strong wood laminates to make highrise buildings.

Bonus trivia: I just ate at Ramin Nagi, a sort of popular Japan noodle house here in the Philippines. I ordered the curry chicken (I can't really stand the ramen noodles, it gives me a stomach ache, maybe I'm getting old or more likely they don't use fresh ingredients here).

Good, the Japanese have the right idea. Fuck McKinsey and their hollow pursuit of labour efficiency, better that everyone has job with decent pay and conditions than every last drop of value get squeezed out of labour's share of income.

I have bad news for you about pay and conditions for Japanese workers

So elderly Japanese, you have a choice: robots or smelly third world peasants?

Yes, that's totally how it works. In other countries the sniffer dogs in the airports are used to search for drugs or bombs. In Japan they are used to find foreigners who don't smell bad enough so they can be barred from entering the country.

At some point in the future people will comment on the strange phase of societal development that included an obsession with having machinery perform every possible task once performed by humans in the name of one of the most important capitalist deities, Efficiency.

Inefficiency tends to get squeezed out for the simple reason that enough of one's customers don't want to pay for it. Revealed preferences are rarely romantic. The Japanese service sector is sufficiently oligopolistic (presumably due to government-imposed barriers to entry) that consumers have no alternative.

One must assume that the equipment used in the production of coffee drinks in modern coffee houses is there because it is more efficient than that used previously. Assuming that to be true, why does a cup of coffee cost multiples of what it did a short time ago even as the price of the coffee beans has stagnated?

The worship of Efficiency is understandable but isn't any more reasonable than the worship of Baal, Jehovah, or the institution of democracy.

Automated checkouts and such are not "robots." It's customers doing the the work that used to be done by store clerks, etc. Good on the Japanese for resisting this (for now.)

I saw automatic check out lanes at Aeon, the largest chain in Japan, around 2010. I just did a 10 second search and found this from the WSJ in 2015:

"Japan’s biggest retail group by revenue, Aeon Co. , has put nearly 3,000 automatic checkers in hundreds of stores nationwide. Wal-Mart Stores Inc.’s Japan subsidiary plans to more than double the number in its stores in 2015."

I'm not sure what the author is talking about. The small grocery store I go to in the U.S. automated four out of six lanes only this past summer.

It looks like Japanese companies are pretty serious about automating check out. From 2017:

"TOKYO – By 2025, Japan’s major c-store chains, including FamilyMart, Ministop and 7-Eleven Japan, will have self-checkout kiosks capable of scanning a basket of items without customers unloading each product, UPI reports. Once the basket is placed on the checkout counter, the kiosk computes each item all at once through RFID tags, rather than clerks scanning the product bar code separately.

A new report finds that FamilyMart, Ministop and Seven-Eleven will have such kiosks up and running in major metropolitan areas next year. These self-checkout kiosks carry a price tag of between $9,000 and $18,000.

Yes, I've seen the automated checkout at Family Mart locations. I think they may have one at the Maruetsu Petit. I use them all the time in the US (I strongly prefer them to dealing with a human at the cash register) but I have never even been tempted to use them in Japan because the experience of dealing with humans at the cash register is much smoother than in the US. I think it will eventually catch on, but there will be resistance. Where I generally see some level of automation is the receptionist at company offices other than showpiece-type headquarters. There, it's often just a phone with a series of numbers to dial for main reception for different departments. Sometimes a touchscreen. It's a huge contrast with the service industry, where the level of overstaffing (vs. a US norm) is dramatic. You see a little of this in places like California (e.g. the people businesses pay to stand at street corners to spin signs around), but it's omnipresent in Japan -- that long line of uniformed employees who bow when the department store doors open in the morning, etc. And, while it's hugely inefficient, I kind of like it. Makes shopping a lot more pleasant.

Recently, I was in an Apple store in Manhattan, and the service was just unbelievably terrible. First, there were probably too few employees to deal with the number of customers. But employees exacerbated my irritation (disgust?) by huddling in a group to chit-chat, while studiously avoiding making eye-contact with customers. My actual purchase took about 2 minutes because I knew exactly what I was buying, but there were 20 minutes first of getting an employee to stop ignoring me and put me in a queue of some sort (which went nowhere), and then just getting a different employee to process my purchase for me. It wasn't quite as bad as Best Buy, but it was pretty pathetic. I've never had an experience that bad in Japan (although I sometimes have to call out or ring a bell in small shops because the owner has left the store unattended).

The inefficiency is not limited to service industries either, I think. I see a lot of maintenance and replacement done on facilities that are in pretty good shape -- e.g. I think they replace the wood decking on some pedestrian overpasses near my condo at least once a year, even though they look like they're still in good shape every time. Hopefully that wood gets reused for something . . .

Visiting from Australia last year, the extensive (over?) use of humans in service roles was notable. 2 - 3 people directing traffic at carpark exits, 3 - 4 people checking tickets / loading bags / standing around on busses, many people standing around at hotels, 2 people at many supermarket checkouts. It feels like it missed the efficiency-driving microeconomic reforms that Australia had in the 90s and there isn't quite as much pressure to cut every possible human worker in the name of cost cutting / meeting competitive pressures. I'm sure it's good at keeping the unemployment rate down and probably does something good for equality.

Efficiency in services is very hard to define. It is easy in manufacturing because the ultimate consumer doesn't interact during the manufacture - so if you use three people to make a bolt rather than two you are automatically less efficient. But lets say you hire a car - if there was only one person on the counter rather than three the efficiency of the rental car company looks high but the person hiring the car has a much lower quality experience. So from a GDP perspective the company that hires three counter staff looks worse - but they may get more customers. Japanese people seem to like quality more than other people, so their measured productivity will look low.
Myself I prefer it when the process is automated and I don't have to interact with someone - fast food restaurants, bank transfers, travel bookings, automated check outs etc are great for me. The learning curve is maybe higher but time gained is huge once you are up to speed. I had to make a bank transfer the other day by talking to someone on the phone as it was above their limits. It took maybe 1/2 hour to do something I can normally do in 5 minutes (and I don't think the security was improved).

People showing you exit from a car parking plce in the middle of nowhere, overstaffing in shops... it smells like a result of some government regulation.

"Japanese customers, especially the elderly, strongly prefer people to machines"... I don't buy that. Are people that cheap in Japan?

Yes, in that Japan has a bureaucratic culture but not one that is uniform or machine-readable, so that constant human guidance is the most efficient way to get around.

Is having people showing you exit from a carpark (that you already know where it is), overstaffed shops...really "most efficient way to get around"? Is marginal productivity really positive for these people? Or is there something I didn't see?

Compared to Europe, Japan seems like an unbelievably wasteful society.

Comments for this post are closed