Recent data shows AI job losses rising, but the numbers don't tell full story
cnbc.com
cnbc.com
Zero real world, specific examples. Just high level numbers and speculation.
AI is actually creating a high number of jobs for consultants and other attention grabbing based professions.
It's CNBC, can't expect much better from them
Right, but I think maybe a lot of the backlash against AI is that no one feels bad or even has much sympathy for these folks. They get to eat, and they’re eating really well. I think people are a little more worried about the Joe and Jane’s who don’t carry computer science degrees.
Or maybe... could these people possibly have been [insert fad here] consultants for years beforehand, and therefore no significant net new jobs have been created so far?
These are specific enough to be credible. If you talk to leaders in customer service / call center operations, they've been planning this for years.
Is this creating some new AI jobs? Sure. But that's a separate issue because peopel who are getting laid off aren't the ones taking up those new jobs.
The layoff has been happening.
Over 240 workers in a call center were laid off this week in KB Bank (a Korean bank) because AI-based consultation tool reduced call center volume by 20%.
Plenty of real life examples happening across the globe.
https://www.washingtonpost.com/technology/2023/10/03/ai-cust...
https://www.washingtonpost.com/technology/2023/06/02/ai-taki...
You do realize some of the most useful BLS and Fed data is little more than surveying some people and asking for their opinions right?
The vast majority of everything we as SW designers do is developed by surveying some users and asking opinions? What do you think user surveys and UX panels that drive so much of SW design is?
It’s a completely legitimate way to get information about something. Like all forms of data it has its benefits and its downsides.
The least capable workers are going to be displaced first and we need to find a way for them to make ends meet. I do not expect the newly-created AI-driven jobs are going to be compatible with their skills. As one industry after another is disrupted, it may prove too difficult for unskilled workers to have to constantly learn new skills, as they jump from one sinking ship to another. This is the present course we are on. There is also the possibility that superintelligence will render wide swathes of the workforce instantly obsolete. If that happens governments will simply have to take drastic action, so things won't be left to the market. But finding a job will only be one of our worries in that case.
I see it like Linux and OSS, of course it killed some paid software products, but it created a platform for companies to build even more and better than before. Devs still have jobs.
AI is a platform, it's even more accessible than OSS, easier to use in any domain and has lower entry barrier. It will support a lot of jobs, and will be the best method to quickly retrain people for their new positions.
What we're basically judging here is if automation and AI will progress faster than our greed. Will we ever say we have enough opportunities and we can stop our efforts now? I don't think so. People will work just like before because we can want more than we can get.
In America, there are no national programs for adult education, most Americans aren't aware of impending industry shifts, and 60% of Americans still live paycheck-to-paycheck without the time/money to adapt themselves.
If AI replaces N repetitive customer service jobs with additional labor demand in engineering, product development, logistics, etc... hopefully new demand = N, but at the same time we really need to ask how practical it is for those N employees to fulfill the new demand. If they can't retrain, the result is some combination of higher wages for those with the skill, offshoring, and lower wages for those displaced into even less skilled labor.
Consider an alternative dynamic: High-skill jobs created by AI will drain the brains from the next lower level skill jobs, which will open up promotion opportunities from even lower skill levels, creating openings at those levels as well. It’s turtles all the way down.
Not saying it’s true, but it could play out that way.
The most automatable jobs aren't perfectly correlated with the least capable workers. Moravec's paradox: what's easy for us is hard for computers, and vice-versa. First job to disappear to silicon? "Computer": https://en.wikipedia.org/wiki/Computer_(occupation)
Right now, ChatGPT is doing a good job of being a free intern-level worker in basically any task that can be performed purely in writing, which kinda-is-kinda-isn't a high bar on the skill level, but it's wild on the breadth of tasks affected.
On the other hand, physical robotics is still hard. Real-time perception is hard. Tesla's gathered a few billion miles of real-world experience, and it's still not good enough to avoid legal battles over the name "autopilot".
I recently installed an AI coding assistant in PyCharm to help myself learn Python better. It's great at completing things for me and once I get used to it, it'll probably speed things up. But ...
I feel like I'm not actually learning the language because I start a line, it finished the function and I look it over to ensure in makes sense, etc., but the actual act of typing out the code and thinking my way thru it as I type ... is gone. I've reduced myself to a QA checker for a computer in California or wherever.
Not thinking about the code == Not learning to code.
Of course, actual government job reports show nothing of the sort.
Lets imagine a fully automated world where 10 billionaires own all the land. For them productivity is pretty much infinity. They push a button and robots make it happen.
But the 6th person that is not part of the owner class is effectively screwed in this world. They don't got no land, and they don't got no robots. They have no means of supporting themselves. The only way they'd survive is by slumming off for scraps of the wealthy.
---
This is the question many of us are asking now. As things become more automated, what value are you going to add to require the ruling class to keep you around?
See https://en.wikipedia.org/wiki/AI_effect - the term "AI" used to include things like chess engines (and even simpler search algorithms), but once computers are able to do those things for a while and people get used to it, the term "AI" tends to shift and only refer to things that machines can't do yet.
It's admittedly a bit of an unsatisfying answer, but whether AI is here is more a question of your definition of "intelligence" than anything else.
Obviously an AI that can have a human like text conversation has broader use than the Go playing AI.
I find it curious that in sci-fi the idea of being human-like was considered the hardest task for an AI, and the walking encyclopedia which is never wrong bit was the easy part.
Seems we did the hard part first with LLMs...
They’re like a 1,000% boost if your job is spamming, advertising for scams, managing astroturf campaigns, or spreading misinformation. So. Hooray.
So for most jobs that aren’t net-negative for society, no, not really, if “there” is “useful enough to seriously shake things up”.
It seems like an easy fit to replace human roles that generate SUBs (Safe, Uninteresting Bloviation) for consumers to read and delete... until the management above these roles realises that it cannot safely outsource its SUBs.
The Marxist response to automation is simply to let the workers have more leisure time.
Still, The Wealth of Nations is 1776 and The Communist Manifesto is 1848, and it shouldn't be a surprise that both look at bit naive with the benefit of hindsight.
Hopefully there's some equivalent work for the impact of AI. Might already exist, I've not actually read "Fully Automated Luxury Communism: A Manifesto", merely some of The Culture novels from whence the title seems to have been borrowed.
But from a positive point of view, there is an entire food industry run by people making food for people. Food will be available. Growing food and bartering for goods and services are common among people. It's just that in the historical economic samples, there was relatively little wealth generated compared to today's economy.
People in the future may have a Basic Income provided by the state. They may spend their time "day-trading", "click-farming", "video-viralising", or "brigading-for-dollars". ^_^