It won't necessarily be a "machine" as such - it may be a computer program instead. Nevertheless, a very large proportion of traditional blue-collar and white-collar jobs are, right now, being replaced by automation; or, at the very least, their replacement is already being planned. The threat is real, and it's immediate, not sometime in the future.
A couple of days ago, I mentioned, in passing, a New York Times article titled "The Hidden Automation Agenda of the Davos Elite". Here's an excerpt. Bold, underlined text is my emphasis.
All over the world, executives are spending billions of dollars to transform their businesses into lean, digitized, highly automated operations. They crave the fat profit margins automation can deliver, and they see A.I. as a golden ticket to savings, perhaps by letting them whittle departments with thousands of workers down to just a few dozen.
“People are looking to achieve very big numbers,” said Mohit Joshi, the president of Infosys, a technology and consulting firm that helps other businesses automate their operations. “Earlier they had incremental, 5 to 10 percent goals in reducing their work force. Now they’re saying, ‘Why can’t we do it with 1 percent of the people we have?’”
. . .
A 2017 survey by Deloitte found that 53 percent of companies had already started to use machines to perform tasks previously done by humans. The figure is expected to climb to 72 percent by next year.
. . .
Kai-Fu Lee, the author of “AI Superpowers” and a longtime technology executive, predicts that artificial intelligence will eliminate 40 percent of the world’s jobs within 15 years.
. . .
For an unvarnished view of how some American leaders talk about automation in private, you have to listen to their counterparts in Asia, who often make no attempt to hide their aims. Terry Gou, the chairman of the Taiwanese electronics manufacturer Foxconn, has said the company plans to replace 80 percent of its workers with robots in the next five to 10 years. Richard Liu, the founder of the Chinese e-commerce company JD.com, said at a business conference last year that “I hope my company would be 100 percent automation someday.”
. . .
There are plenty of stories of successful reskilling — optimists often cite a program in Kentucky that trained a small group of former coal miners to become computer programmers — but there is little evidence that it works at scale. A report by the World Economic Forum this month estimated that of the 1.37 million workers who are projected to be fully displaced by automation in the next decade, only one in four can be profitably reskilled by private-sector programs. The rest, presumably, will need to fend for themselves or rely on government assistance.
There's more at the link.
If you think that article is unduly alarmist, think again. Here are a few headlines from the past week or so. Follow the links to read them for yourself.
- Citi Ready To Replace "Tens Of Thousands" Of Call-Center Workers With Robots (if the link leads to a paywall, see the summary here). Money quote: "Under pressure to bring its cost base in line with peers, Citi executives have been upfront about the impact of technology on their 209,000-strong global workforce, including last summer’s warning that as many as half of the 20,000 operations staff in its investment bank could be supplanted by machines."
- The Most Mindnumbing of Office Tasks Made One Man $360 Million. "Takahashi’s firm provides so-called software bots for more than 500 companies ... It helps them to automate routine tasks such as inputting data and checking invoices ... “There’s a huge market for software robots using AI technologies,” he said. “Just like industrial robots in factories, if software bots can take on the tedious routine work in offices, we can create a productivity revolution for white-collar jobs."
- OpenAI's new multitalented AI writes, translates and slanders. "OpenAI’s new algorithm, named GPT-2 ... excels at a task known as language modeling, which tests a program’s ability to predict the next word in a given sentence. Give it a fake headline, and it’ll write the rest of the article, complete with fake quotations and statistics. Feed it the first line of a short story, and it’ll tell you what happens to your character next. It can even write fan fiction, given the right prompt." (Not a comforting thought for me, as a writer!)
- China has produced another study showing the potential of AI in medical diagnosis. "Researchers found that the AI system was able to meet or outperform two groups of junior physicians in accurately diagnosing a range of ailments, from asthma and pneumonia, to sinusitis and mouth-related diseases. The AI was also able to meet or exceed diagnostic performance with some groups of senior physicians, for instance, in the category of upper respiratory issues ... In no category did the AI model dip below a diagnostic accuracy rate of about 79%..."
It used to be said that only repetitive office tasks such as basic book-keeping, order entry, etc. were at risk of automation in the short term. However, as the links above show, that now extends deep into white-collar territory. It also has implications for supervisory and management positions. If the junior staff aren't there, they don't need supervisors or managers, do they? Some expertise will have to be retained in-house, but I'll be surprised if it's as many as one in four of the numbers who used to be employed in such tasks. I think it'll be more likely to be one-in-six, perhaps one-in-eight - and the ratio will decrease even further over time.
If you're in any job that can be handled by an automated system, whether blue-collar or white-collar, you need to be looking into re-training and re-skilling yourself, right now. It's as critical as that. It can no longer wait for some nebulous future date - and if you do wait, you'll be part of a flood of newly unemployed people, all trying to do the same thing. The competition will be intense, perhaps ruinous. Far better to beat the rush and start the process now, even (if necessary) at the cost of some short-term monetary pain. If you work for a company such as Amazon.com, with its Career Choice program, your employer may pay most of the costs: if not, sadly, you'll have to fund them yourself - but it remains a worthwhile and necessary investment. Train for a job where you can make yourself indispensable, or where a machine or AI replacement will be difficult to accomplish. (The skilled trades look very good right now - plumbing, HVAC, vehicle repair, etc. are unlikely to be automated anytime soon, due to non-standard conditions in the field to which a robot or AI system will find it hard to adjust. However, I daresay their time will come.)
As for me, I'd better write better! If an AI can produce "formulaic" fiction good enough to sell, I'd better make sure I can write to a higher standard than that, so I can continue to make a living. Forewarned is forearmed!
Peter
And yet, disabled black gay trans women will still "need" to be hired by every company in the country to meet mandatory diversity quotas. Whether there is anything for them to do, or not.
ReplyDeleteTo understand the US federal government, you have to understand that two-thirds of the workers in DC are black women. Match that up with the fact that only on in six 8th graders in DC will go on to graduate from high school able to read and write.
But to fix that bust sewer will still require a plumber...
ReplyDeleteTwo questions stuck as I read your piece...
ReplyDeleteIf robots are doing the work who's going to buy the product? What do the "Davos elite" need from the masses?
You mentioned the writing software, it's not just the written word.
I was watching a youtube video the other day & after a couple of minutes I stared noticing something a bit odd about the narrator, I think it was a MUCH improved computer doing the talking! Then I wondered who made the video and what their agenda was.... This modern world it does not have to be an English speaker anymore.
Keep in mind that robots don't need to be better than humans, they just need to make fewer mistakes...
The Slate Star Codex blog has been talking about GPT-2 the last few days. You might find it interesting and perhaps reassuring.
ReplyDeleteNot so very long ago there was a major adjustment to the outsourcing of call center work as companies found that their customers balked at trying to navigate through levels of english speaking Indians, for a very liberal definition of english speaking.
ReplyDeleteIt's important enough that some companies use US based customer service in their advertising.
The other thing is that I've watched for most of 60 years the migration of most manufacturing jobs overseas. The stated reason was always cheap labor, though I suspect an added benefit was fewer restrictions on the manufacturing process due to government regulations, EPA, OSHA, and that whole alphabet soup.
But when the entire process is mechanized, the human factor all but eliminated, the not insignificant cost of transportation has to become a factor as well. Let the foreign producers automate their factories. Eventually they will likely realize that moving production closer to demand is a financial necessity, much as has already been done with several car manufacturers. They are in the process of building a $9 billion combined Toyota/Mazda auto plant even as we speak here in northern Alabama.
I'm an automation engineer and I don't buy a lot of this hype over automation and AI. Automation has been on-going for the past 50 years and has been progressing in an incremental manner. This is not changing. A lot of blue collar work has already been automated over the past 30 years. A factory that employed 500 people in 1970 now employes maybe 50 people with many of those being maintenance technicians. Good maintenance, hydraulic, and electrical people will never be out of a job.
ReplyDeleteI think the automation of white-collar work will be more dramatic over the next decade or so. That's because a lot of the automation that could have occured since the 1980's didn't, for reasons that are not clear to me. I suspect office politics and the desire for managers to have lots of underlings and to build their own little empires had a lot to do with it.
The thing is that it is often easier to automate white collar work than blue collar work because the white collar work is all with computers these days. Blue collar work is physical and its difficult to make robots that can do complicated tasks (like wire an electrical panel). Since the blue collar world has already automated to a large extent, blue collar jobs today may be more secure that many white collar jobs.
So instead of getting a non-helpful tech support answer from someone in a third world country with poor english language skills, you can get your non-helpful tech support answer from a robot that mimics perfect english.
ReplyDeleteMakes me glad to be retired.
ReplyDeleteThe sky is falling, the sky is falling!
ReplyDeleteContrary to the 2017 Deloitte survey, 100% of companies use machines to perform tasks perviously done by humans. You just have to go back to what humans were doing 250 years ago (e.g. all writing, calculating, etc).
Progress and automation have been going on for centuries, we haven't seen mass unemployment, and our standard of living continues to improve. I'm going from memory (can't give references) but I've been hearing "all the jobs will disappear" or "Japan/China/whoever" will take all our jobs for 40 years. The computer industry was supposed to disappear around 1990 after Japan's 5th generation computing project. Expert systems were going to replace doctors in the 1980s. The list goes on but hasn't come to pass.
The last couple years we were told economic and job growth would be permanently slow to stagnant, yet jobs are being added to the economy in leaps and bounds.
But still, the sky is falling!
The silver lining to AI diagnosis be it medical or automotive troubleshooting is that once challenged people get better and people with AI assistance get much better.
ReplyDeleteThe Apartment is a great movie but in big buildings automated elevators with predictive technology move more people faster and IMHO leaving fewer people at the very bottom is a plus.
As someone pointed out above, at what point have you "automated" so many people out of work that there are no customers to buy the products or services that produce the income? Many jobs can be taken over by machines/software but there will be a point of diminishing returns.
ReplyDeleteTo illustrate, an acquaintance toured the Peoples Republic of China in the 1990s. He saw a large number of men breaking up rock into gravel with manual tools. He asked the tour guide why they did not use a gravel crusher. The answer was essentially that they needed to keep people employed so they did not use such machines.
The world that some of these people dream of that makes them more money by getting rid of workers is just not functional in reality.
I'm very concerned about middle managers and office workers being increasingly automated away.
ReplyDeleteLike when a VP can directly access a data cube of all the sales & other revenue (in Tableau, for instance) and, in a meeting, look at it by year, by quarter, by week; by product; by sales team (salesperson); by region.
Then, if there are questions, the VP can ask the computer assistant to get additional info from the various, not-always-consistent databases.
Lots of current middle layers are at risk in the coming 2020s decade.
We need policies that support the creation of human jobs in America, as well as less taxes on the normal folk (meaning less gov't spending or more taxes on the rich).
A carbon tax that increases transportation costs, yet is refunded equally to all US workers, is one policy that might help along these lines. We need society, thru elected politicians, to be trying different things out.
Automation is already restricting us to "ten choices" as it is.
ReplyDeleteIf your issue is something arbitrary and idiomatic, forget any communications with an A.I. or automated service as your particular issue won't fit any "modes' programmed into the system.
And don't forget: if you move or change your phone number A.I. will be sending your mail to two addresses and calling you via two phone numbers---because all it ever goes by is clinical data.