Artificial intelligence will destroy 'laptop class' workers

[riversedge]

Change is always coming--but this article makes it seem it is coming VERY fast.


Artificial intelligence will destroy ‘laptop class’ workers

https://thehill.com/opinion/technology/3884324-artificial-intelligence-will-destroy-laptop-class-workers/

by Kristin Tate, Opinion Contributor - 03/08/23 10:30 AM ET


The coming artificial intelligence economic revolution will be a major shock to the world. There is a serious possibility that the next decade will bring about a series of social and economic changes akin to the Industrial Revolution and the advent of the internet combined. Many writers, human resource officers, lawyers, writers, artists, and even coders increasingly will be replaced by AI as the “laptop class” of workers is decimated. At the same time, blue-collar workers who work with their hands will enjoy job security; their services cannot be replaced by technology. Unfortunately for waves of young people, the media’s advice to “learn to code” may have been like investing in typewriters.

Artificial intelligence is advancing at a breakneck speed. Recent announcements of programs that can mimic human conversation, copy our voice, write research papers, and paint beautiful pictures are just a small sliver of the coming AI revolution. The coming changes in everyday life soon will become noticeable, including the popularity of AI-generated video games, music, art, and even movies. A short description and a click of the mouse can spit out a new novel by John Steinbeck or an economic treatise by Thomas Sowell.

Scores of jobs that require a college education will be changed nearly overnight. Rapid advances in this new technology will wreak havoc on the very people who prospered during COVID, especially those who work in the “knowledge economy” and can often carry out their duties from their laptops at home. Artificial intelligence advances within the next one to five years will outpace most work a human can input into a keyboard. Most content on the web will be written by chatbots. There will be AI influencers. Code will be written in a tiny fraction of the time it takes for humans to produce it. Graphic artists will lose most of their business to art generators. Even accountants and financial analysts may be outpaced by computers. ChatGPT already helps coders through basic code, which often needs refining. The chat service also can help replace many of the smarts needed to build a website. It already has passed an MBA exam and law exams.

Some white-collar jobs will fare better than others with the advancement of AI. Those who pioneer new techniques or are at the top of their fields will still be able to earn a respectable wage. At the same time, workers whose jobs rely on an element of personality and face-to-face relationships likely will weather the storm. Would you rather have an actual human advising you on legal matters, your finances, and your health care decisions — or a machine? Still, the shift toward a nearly labor-free creative world may make white-collar jobs across the board fewer and farther between.
.............................

[moosewhisperer]

1. She needs an AI proofreader

“Many writers, human resource officers, lawyers, writers, artists…”

Writers twice in one sentence? She definitely should be worried about being replaced by a robot.

[Orrex]

7. That jumped out at me as well

I'm sympathetic, but it's hard to make the case that you're essential when you commit such a basic and ill-timed goof in the process.

[Celerity]

38. she goes by 'The Libertarian Chick' lol

some of her 'deep' claptrap from 2016

The Libertarian Chick on Government Gone Wild!

[Auggie]

2. This sucks ...

said a human.

[Bernardo de La Paz]

21. In same way as steel making facilitated kitchen knives and swords simultaneously

1) Article is far from the whole story and is a bit glib.

2) "Sucks"? Sure. The steelmaking that made swords sucks, but the same knowledge makes outstanding kitchen knives, which does not suck.

So "sucks" is exceedingly shallow as a reaction and I think you need to dig deeper or at least avoid the failure of excessive terseness, since I doubt you are that shallow.

[MayReasonRule]

3. I'm not convinced her arguments are substantive.

A very interesting opinion piece though!

Thanks for the post!

[Amishman]

31. She has good points, but overstated

I work in software development as an analyst / project manager.

machine learning 'AI' can spit out useable code - with a few caveats. It does better with simple or common functions, but for stuff like that there usually is open source code available that a smart developer would be leveraging anyway. It also needs to be tailored and fitted to your specific system and data structure. It also needs significant testing - if anything more than human written code. So it's a nice little shortcut to make developers more efficient, not a true replacement.

Technical writing is the same, but with less massaging to get the AI output useable. The fewer hard facts / details in the subject matter, the more AI can help.

It's best at writing 'fluff'.

And that is another spot where it falls short. The article says it can spit out a new Steinbeck novel or an Sowell economic paper.

The former it can just about do today, fiction and regular prose is something it does well. It can blend elements countless different existing stories and ideas so that a reader doesn't recognize the source materials, then shape it with a specific author's preferred word choice, syntax, and style.

The economic paper is trickier - it won't have any coherent new ideas. You will end up at best with something rehashing and discussing old ideas that are in its data model, or original content that makes no sense to someone with a solid understanding of the subject matter. So it can do a good job writing a high school or undergraduate student's kind of paper discussing the existing ideas of Sowell, but it can't give you new original research. It simply doesn't have the depth of real world information needed to come up with something original that makes sense against a real world backdrop - let alone the computing capacity to parse a data model that size if it could be created.

The white collar jobs that really are vulnerable are people like accountants and underwriters - those who work with numbers, repetitively applying standard and well defined rules to data.

[multigraincracker]

4. Change is the only thing

that remains the same.

[modrepub]

5. Creative Destruction

as Shumpeter (the only Austrian economist I respect) said. The economy is always looking for new (cheaper or more efficient) ways of doing things.

Change is the only constant. It's always good advice to update your skill set and or try new things if you're going to participate in the market place.

[Tanuki]

6. Consider the source. The author, Kristin Tate:

"Kristin Tate is a libertarian writer and an analyst for Young Americans for Liberty. She is an author whose latest book is “How Do I Tax Thee? A Field Guide to the Great American Rip-Off.”

[Tanuki]

9. https://en.wikipedia.org/wiki/Young_Americans_for_Liberty

..."In total, YAL has fought COVID-19 vaccine and mask mandates on 23 college campuses, such as Rutgers, Virginia Tech University, and the University of Colorado Boulder.[39][40] YAL has circulated petitions opposing COVID-related rules on different campuses, such as Virginia Tech.[40] According to the YAL, the organization "is not anti-vaccine, but rather anti-vaccine mandate at taxpayer-funded academic institutions."[39] Several students affiliated with YAL have spoken to the mainstream media, arguing for vaccination as a "personal choice."[41][42]

YAL is an outspoken critic of gun-control legislation, such as "red flag" laws proposed by President Biden and other Democrats.[43][44] The organization also opposes "Critical Race Theory" education at public schools.[45] YAL is active on social media, often attacking Democratic Party officials.[46] The organization also publishes a quarterly magazine called "The American Revolution."[47] "

[Bernardo de La Paz]

17. Unspecific assertions smear people. Guilt by association. What in the ARTICLE do you object to?

Be specific, please.

[hunter]

27. Anyone who writes libertarian drivel is already a bullshit machine.

Of course they worry about their jobs being automated.

[Ford_Prefect]

8. Kurt Vonnegut was right! His first novel, Player Piano, is about to become reality...

Player Piano is the first novel by American writer Kurt Vonnegut Jr., published in 1952. The novel depicts a dystopia of automation partly inspired by the author's time working at General Electric, describing the negative impact technology can have on quality of life.[2] The story takes place in a near-future society that is almost totally mechanized, eliminating the need for human laborers. The widespread mechanization creates conflict between the wealthy upper class, the engineers and managers, who keep society running, and the lower class, whose skills and purpose in society have been replaced by machines.

https://en.wikipedia.org/wiki/Player_Piano_(novel)

[ananda]

12. The prescient futurist writers have always impressed me.

Stanislaus Lem is another good one.

His novel about virtual reality was something.

And Philip K Dick's works are also great.

[FSogol]

48. +1 Exactly. n/t

[Bernardo de La Paz]

11. The AI revolution is over-rated and under-rated at the same time

Advancements like AI go in waves. The writer is surfing a wave of over-rating AI, but is helping wake people up about AI and the people need waking.

1) There have been waves of AI before. https://en.wikipedia.org/wiki/AI_winter

2) AI is a very real human capability, by which I mean it is within our grasp. But it is not a thing which arrives or has not arrives. AI capability and AI deployment come in degrees, levels, etc.

3) Since about 2000, long term research threads began to bear fruit as artificial neural networks (ANNs or NNs) became more possible with advanced computing hardware and cheapness of generally available hardware.

4) In the 2000s, large "tech" companies (meaning web companies) began adopting and integrating NNs in massive ways. The emergence of Big Data. It was the business model of Google, Facebook, and the rest: your data. Their joint agreements mean they all share data since the ToAs of them all say that you consent to sharing of data with joint venture partners.

5) The advancements of AI currently are stunning. No question. That is a big part of why I say it is "under-rated".

6) The advancements of AI currently are flawed. That's why I say it is "over-rated". The article is a bit glib and too ready to predict a calamitous swift revolution. It will be swift enough to amaze us, but it will go in fits and starts.

7) Flaws are hard to root out. But they will be. It will take time. This is why I think that the AI industry will head into a bit of a winter before long.

More on this point: Sure, an AI can write a bit of code, like implementing known algorithms in toy applications. But to write a complete working 20,000 line program involves designing and writing code when the problem is not well defined and gets defined by writing working code that fails to satisfy users then modifying it in steps until it does.

If you can't define it precisely, then the AI can fake it, but not make it.

So what will happen is that there will be AIs that facilitate interaction with AIs. Software will be developed in conjunction with engineers and programmers, but their mode of interaction will be higher level than it is now.

Developing the software for those interactions will take time. Hence AI is over-rated. But AI will assist that development. All up and down the line AI will be accelerating those developments and at some point in the future it will seem like the "next wave" has suddenly arrived. Hence AI is under-rated.

This will happen across all the fields. AI integration will take time but will be amazing.

Take artwork for example. You can pick an artist known for a fairly consistent style and ask the AI to make a picture of something similar to their themes in their style, and the result will seem perfect at first look. But after a bit we notice flaws like a car poking through a window. This will all be ironed out, but it takes time.

The graphic artist of the future will do some "painting", but a lot of their work will be interacting with an AI. When they tell the AI "looks like a car poking through the window and that's nonsense", the AI will correct it and there will be back and forth.

So. Advancement was ever thus, in all spheres of human endeavor.

[highplainsdem]

41. A new AI winter can't come soon enough, IMO. I'm hoping for both

a flood of successful lawsuits shutting down the companies whose datasets have ripped off intellectual property, and publicizing of business failures and security breaches that will slam the brakes on this use of AI before almost everyone involved gets hurt except the tiny percentage raking in profits from the new AI stampede.

Even the people behind the stampede admit government oversight and regulation are needed. But they haven't been willing on their own to give governments and laws any chance to catch up.

[Bernardo de La Paz]

47. Lawsuits are the bane of libertarians' existences since they hope to escape liability

It's an odd kind of sense of entitlement, and the Achilles Heel of the magatized version of libertarianism as promulgated by the rump and the mollusk.

Data is the tip of the iceberg as far as the need for regulation and oversight.

Getting it right is tricky, but getting started would be a good idea.

[Jazz Jon]

13. Exagerated

Not as soon or as dramatic as the writer believes. I have encountered "AI" in three areas.

Art creation and editing tools... they speed my work, but don't eliminate it. They have no understanding of their task.

Telephone sales... really bad so far. They can't respond accurately to the first question.

Customer service chat bots. Fairly bad... They can react to simple sentences, but don't comprehend meaning. They merely return short entries from a "knowledge base" that have matching keywords.

[Jazz Jon]

15. Chatbots are obstacles

If you are trying to get help from a company who's services you have bought, and you encounter a chatbot, it's a major headache. You have to figure out how to get around the chatbot in order to get your question answered or a get an action taken to solve your probllem.

[Amishman]

33. call center / help desk chatbots are based on defined scripts

I am not aware of any that leverage a deep machine learning model to give better answers that are not explicitly scripted.

Why are companies doing this?

Because there are not nearly enough controls in machine learning chatbots like ChatGPT to ensure the bot doesn't promise something that they company is unwilling to give. We see this with examples where 'AI' chatbots literally make up sources and present their lie as fact. Now imagine a bot bugging out like that when a customer is asking about their bill or mortgage payment. A bank can't risk having their bot glitching and telling a customer they don't need to make their payment.

[Bernardo de La Paz]

49. The first level of AI usage in call centers is natural language recognition; there it's tougher

The use of natural language recognition in centers is tougher than, for example Alexa, for multiple reasons:

* the variety of inputs is extremely wide and deep: accents, age, cultural patterns, sex, health, strength, socialization, trends like vocal rise and vocal fry, emotional turmoil, etc.

* there is no training with the individual speaker

But still, it does give them leverage.

[Bernardo de La Paz]

50. As to bots bugging out, AI will not replace traditional corporate software

It will infiltrate, or rather the engineers will infiltrate it in to the systems in place. Some systems may be redesigned for other reasons and when that happens, systems and subsystems will be rebuilt with more integration.

But regardless, there will be databases that can be consulted, and audited, and logged as they are now. AIs will use them and humans will use them, most often the same one(s).

So the bank would limit the AI bot from certain behaviors or require certain behaviors such as it must communicate the exact amount from standard SQL database queries that are logged.

Even so, there will be glitches and mistakes and wild media coverage and lawsuits and payouts.

[Bernardo de La Paz]

20. The AI of today is not the AI of the next decade or the one after that. Do not be lulled. . . nt

[Celerity]

35. I just wanted to say your contributions on this thread are admirable heavy lifting, and appreciated

[Bernardo de La Paz]

39. Coming from you, Celerity, I deem that high praise. Thank you

I like seeing your name among the posters and often head there first.

[Celerity]

42. aww, thankies

[BannonsLiver]

40. Hate and fear of AI is now customary on DU

[Trueblue Texan]

16. I think the fear of AI is a lot of hype...

...AI writing is merely regurgitation with good sentence structure. It uses info from the internet so of course, it must be deeply fact checked. I think AI will make real writers in greater demand because, yes, there will be more written works due to the easy accessibility of AI. But there will be far more need for fact checkers and logic checkers. AI that can truly be intelligent is a long, long way off.

[Blues Heron]

18. A short description and a click of the mouse can spit out a new novel by John Steinbeck

yeah, no.

[Midwestern Democrat]

19. The author lost me when she stated that AI will be able to produce a novel at the level of a John

Steinbeck. Give me a break - I've seen computer generated web content and You Tube videos - they're awful. I've seen You Tube videos generated by computer where they mispronounce a famous actor's name; I saw one on an actress where they included photos of an entirely different unknown woman who just happened to have the same name. AI today might be able to comb through Wikipedia pages, etc. and produce stilted, boring narrative, but I've seen no evidence it can come anywhere close to duplicating a human expert's ability to make value judgments of what's really important and interesting.

[Bernardo de La Paz]

24. It would be a mistake to judge the prospects of AI, the same mistake as

It would be a mistake to judge the prospects of AI by its current state, as much as it would have been to judge the prospects of "cell phones" by those available in 1993.

Same mistake as to judge AI by the state of the field in 1993, or in 2023.

I am likely to be dead in 2053, but if you or I live even 15 years more, we'll be blown away by applications of AI in 2038. It would have been a mistake to judge the prospects of self-driving cars by their capabilities in 2008.

I'm not expecting a Steinbeck novel by 2038 or 2053, but I expect to be as amazed by the advances in 2038 and 2053, or however close I get to those, ... as amazed as I have been by advances in the last 15 years.

[bucolic_frolic]

22. The self-driving car moves into language and professions

There will be crashes. Some of them will be embarrassing. Some will cost a lot of money to settle. But the top of the pyramid will continue to be fed well.

Didn't someone write the AI code to begin with?

I've been reading for several years about trading bots running ETF's more profitably than humans. A few are said to have no human oversight. But, again, someone wrote the algorithm.

[Bernardo de La Paz]

25. This is the real thing: "fed well". There is one thing that should come out:

If the AI revolution is going to make some people enormous wealth (it will), it will not be without risks for society.

One big risk is further widening the unsustainable wealth and income gap.

That must be closed by providing basic human necessities for free: housing (a room with basics like shower), health care, food. No questions asked.

If the AI revolution is good for anything, then all people must be fed well and "fed well".

That must be paid for by higher taxes on the wealthy. It is a myth that they did it all themselves without infrastructure like roads and government services like public education.

The result will lift all boats and even the wealthy would be happy. Well, mostly happy. They would still moan and groan about taxes at their yacht parties.

[bucolic_frolic]

26. Defining legal superstructure is what our Founders did, & someone still has to do physical work

This is analogous. Slavery enters its next rendition. We are ant or bee colonies.

[hunter]

28. I'd add free education to that list as well, at all levels...

... from basic literacy and numeracy to doctorates.

Our civilization is productive enough to support that.

[Bernardo de La Paz]

37. +1. And the payoff to all people is so big from education. . . nt

[Silent3]

36. Actually "someone wrote the algorithm" isn't true with self-learning systems

The only thing that humans have written are the algorithms for how to go about learning by analyzing tons of data, and how, by trial and error, to find some way to match inputs with expected outputs.

Those "some way" solutions are generally incomprehensible to the humans who created the AI. All we know is how well they perform, and we're quite often surprised by the mysterious ways they can fail.

[Javaman]

23. when I was promoted to my current job 8 years ago...

six months in to the position, I turned to my coworker and said, " you know, in a few years someone is going to write an algorithm, that will replace us."

I still believe that. I just want to hang on long enough till I retire. just 7 more long years.

[Happy Hoosier]

29. Have a job that requires creativity and judgement.

Or a skill that cannot (yet) be automated.

Pretty much, if your job doesn't require exercising judgment to make decisions or actual creativity, your job is a candidate for automation.

And yes, Chat bots and AI "art" can emulate SOME creative endevours, but they are derivative, not creative, so they can be used from some basic elements, but they cannot replace creativity. Not yet.

So far, the AI writing is still reasonably weak, but it will get better. Especially when it comes to subjects requiring actual analysis, the writing is a word salad. Of course, the writing of a lot of HUMANS is a word salad, so...

But for now, the "AI" algorithms cannot exercise actual critical thinking or genuine creativity.

But if your job is to put peg A into slot B, or any kind of "data entry" yeah... it's just a matter of when it becomes cheaper to have a robot do it.

[thesquanderer]

30. AI may replace commoditized art and music "in the style of" but there will always be a place...

...for genuine artists. The problem isn't that AI will take away their livings, it's that so few of them can make good livings from their art in the first place.

[hunter]

43. Society would still progress without superstar artists, actors, athletes, etc..

I'd argue that any artist who gets trapped in some crap job related to their art is LESS likely to pursue their actual art.

Most of those kids hoping to be superstar football or basketball players are likely to have their dreams shattered. Same with any other sort of artist.

Ideally as an artist you find the means to pursue your art without compromise. You don't get assimilated into money making machines run by asshole billionaires who, in any sane civilization would be taxed out of existence or possibly placed in prison to protect the rest of us from their predation.

If we weren't all clueless fucks we'd demand thirty hour work weeks so we could all pursue our own arts. Our 21st century technology has given us all the tools we need to create that kind of society, we just have to apply them.

As it is, most of us suffer work that is not making the world a better or brighter place, and for no good reason. We live in a society that stifles curiosity and creativity.

That's one of the reasons I don't support advertising supported television. I simply don't see any television advertising in my daily life. No broadcast, no cable, no satellite, no streaming with ads. It no longer exists within my personal universe. I've seen too many artists sucked into that wormhole who never came out the other side.

I tend to think any job that can be automated should be automated. The problem has always been that the benefits of this automation are not distributed equally among us. The wealth doesn't "trickle down" to those whose jobs have been automated. In a better world things like comfortable basic housing, healthy food, appropriate medical care, and education wouldn't require anyone to accept soul-crushing work in exchange for mere survival.

[Sympthsical]

32. I think a lot of media people will lose their jobs, and that's why we get these articles

Because anything that happens to the media has to be discussed into the ground and flagged as Very Important Thing.

I've been reading more and more articles online that follow kind of the same format. Very click-baity headline designed to draw attention with a characterization meant to invoke an emotional response. Then maybe two or three sentences kind of outlining what's being discussed.

Then paragraph after paragraph of background information that may or may not be relevant but definitely feels like it was copy and pasted from somewhere else.

Towards the bottom: Maybe two sentences detailing what actually happened, relevant information, or some throwaway bit that actually invalidates most of what just came before if not contradicting the headline outright.

Humans are not required to write this stuff. I think online media writing is where a lot of low-achieving English majors ended up being employed.

Places like Buzzfeed and half of what's on MSN already feel like mindless AI is writing it. They'll be the first to go.

"AI is coming for all our jobs!" No, dear. Just your job. The rest of us have useful skills.

[Silent3]

34. This change is definitely coming, but how fast is very uncertain

ChatGPT is simultaneously very impressive in some ways and deeply flawed in others. The flaws won't be truly fixed until some major breakthroughs in "general AI" (as opposed to "narrow AI" ) are made.

In the meantime, problems with ChatGPT-like systems can probably be whittled away at a bit, masking them or burying them a bit deeper.

We already seem some companies treating ChatGPT output as "good enough", but, like Fox News, they must care more about money than accuracy and factual truth.

[sarcasmo]

44. Universal basic income, now!

[Bernardo de La Paz]

51. +1. . . .nt

[Elessar Zappa]

45. Society and technology changes.

As I’ve said on other posts, there will be good and bad effects of AI. Certain laws regulating it may help but I’m not sure what those laws would entail. Ultimately though, I’m not losing any sleep over it.

[RobinA]

46. I Work In Mental Health

AI, bring it on. I'm not scared of you!

[honest.abe]

52. I dont think so.

The "laptop class" workers will simply get better at doing their "laptop class" work.

[Torchlight]

53. It'll be interesting to see how it eventually works itself out.

I'm a guy who claimed the internet would never be close to ma bell's wall phones in popularity, so I know my guesses are more often wrong than not, and I won't even try to pretend I have a gauge on its eventual usage, whether for good or no.

It's weird to me (as an adult having watched a few waves of tech progress crash onto the shores) how the world can adapt to newer tech so quickly and (seemingly) effortlessly, allowing what fascinated me only a year ago as the "it just can't get any greater/faster/smaller/bigger than this!" to become yesterday's jam and ubiquitous today.

[XorXor]

54. Hyperbolic sensationalism

Had ChatGTP not been made so accessible to the regular public, then the vast majority of these articles that are written by non-technical people wouldn't be a thing. Yes, they are cool, fun, and can be useful, but anyone who does more than just play with them briefly will come to understand their limitations.

Maybe there will be a massive disruption in the future, but we are not there now.