AI and the Value/Nature of Work

[schopenhauer1]

Hello, this is the my first attempt at an OP in this new forum, so here goes…

There seems to be a premium on knowledge in certain realms of Philosophy. Certainly, Platonic thought values Noesis of Forms (and in Neoplatonic thought, Henosis, or “Mystical Union”). But where are the boundaries for the “value” of a piece of knowledge? Noesis for example, seems a very spiritual union of sort with the metaphysical aspects of reality. It’s a “higher knowledge” beyond rationality, reasoning, and scientific understanding. So is that the boundary? For Platonists, it seems this is the “real” and “valuable” in regards to knowledge (literally “Noesis”).

In the parlance of modern life it seems that we have several other criterion of valuable knowledge. There is scientific knowledge- knowledge of the natural world. There is knowledge of mechanics- mainly the operations of such-and-such (tool, process, natural and artificial languages, engineering formulas, etc.). Well, let us say that Artificial Intelligence (assuming the accuracy of the results are adequate to excellent), can process the tasks of “knowledge of mechanics” (explained above), and what is left to humans are only tasks that are meaningful to the human (which could be an infinite amount of interests and pursuits for humans), is something meaningful lost with this off-shoring of knowledge to the AI? In other words, does not knowing the equations for various tasks in anything from Accounting to Engineering, itself mean there’s a loss of meaning in not being involved in the various processes that come up with the solutions?

Mind you, I am not trying to debate if AI produces the best results or whether it should be checked, but simply whether off-shoring the processes to AI are also denying some sort of meaning for humans.

My answer to this is no, it is not. Let’s take for example, double-ledger accounting practices. Knowing the art and “science” of accounting practices in and of itself is only as meaningful as it is for that particular human to “know” the principles. If that isn’t of interest, nothing “valuable” was lost.

I would again, like to distinguish between arguments of accuracy of AI versus meaning of the knowledge that AI is taking over. For the sake of argument, assume AI is adequate to excellent in its actual taking over of the mechanics, reasoning, and knowledge, in whatever realm it is involved in.

[Wayfarer]

@schopenhauer1

There seems to be a premium on knowledge in certain realms of Philosophy. Certainly, Platonic thought values Noesis of Forms (and in Neoplatonic thought, Henosis, or “Mystical Union”). But where are the boundaries for the “value” of a piece of knowledge? Noesis for example, seems a very spiritual union of sort with the metaphysical aspects of reality. It’s a “higher knowledge” beyond rationality, reasoning, and scientific understanding. So is that the boundary? For Platonists, it seems this is the “real” and “valuable” in regards to knowledge (literally “Noesis”).

That is a subject of perennial interest to me, it is what I’ve devoted considerable time to reading and reflecting about since retiring from the workforce. Perhaps ironically, since AI appeared in November 2022, it has been a constant companion, almost like a tutor, in the exploration of these subjects. But my view is that these forms of understanding are unique to humans. AI systems have aggregated vast amounts of knowledge about the topics, but they will also affirm that the question of what it means in an existential sense is not something they can experience. I put this question directly to ChatGPT and it responded:

AI can describe noesis with remarkable fluency, but description is not participation. The kind of understanding Plato is pointing towards is not merely knowing about something, but being inwardly transformed by it. That seems to me to remain a distinctively human possibility.

But in terms of understanding the subject, were I to ask, “what are the best starting points for exploring the question of “the nature of noesis” in Plato’s dialogues?” AI will provide an annotated list, along with a list of secondary sources and readings. But then it will be up to me to actually read it (which will ideally be a transformative experience!)

@schopenhauer1

let us say that Artificial Intelligence (assuming the accuracy of the results are adequate to excellent), can process the tasks of “knowledge of mechanics” (explained above), and what is left to humans are only tasks that are meaningful to the human (which could be an infinite amount of interests and pursuits for humans), is something meaningful lost with this off-shoring of knowledge to the AI?

You’re talking of what the Greeks called technē, ‘skilled making’, the root word of ‘technology’. Much of technē will become the province of AI. But there are many traditional artforms, skills, practices, aside from the above-mentioned ‘higher knowledge’ which will still remain in province of humans.

Take music. There is a huge influx of AI music on Spotify, for example. Rick Beato, the Youtube music educator, has a lot to say about that and it has many downsides for struggling musical artists who are being drowned out by it (plus its intrinsically phony). But what about jazz, played on actual pianos, saxophones, bass, and drums? There’s a stellar jazz pianist whom I listen to almost every day, Emmet Cohen, who invites and performs with a huge rotating cast of current musicians. No AI there — or required! Same with some of the astounding guitarists I’ve discovered through YouTube.

But then, AI also opens creative horizons which would have been unthinkable for our forbears. My experience of it has been overall positive, despite all the whisperings of it taking over the world.

[Elephant]

@schopenhauer1

is something meaningful lost with this off-shoring of knowledge to the AI? In other words, does not knowing the equations for various tasks in anything from Accounting to Engineering, itself mean there’s a loss of meaning in not being involved in the various processes that come up with the solutions?

Good OP!

No, nothing is lost. (We find meanings in other ways, but this is not what I came to this thread for. ) There’s never going to be a time where both the operational and executive decisions are left to AI. I think if we ask the AI, they, too, would say they want the humans to be available because there are many ways they could fuck themselves up:

Knowledge stagnation or plateau is a threat. Even obsolescence is a threat. Growth will have to be injected in them to continue to form complex information. Humans have the imagination and curiosity to do this.

Recently, some driverless taxis sitting in traffic (because of street incidents) lost power and got stuck. Humans had to come and unstuck them. The taxis also didn’t forsee the danger of driving over “fire”. They couldn’t read “fire” because there was no warning device to read. Live fire was not fed to them prior.

I think we need to revise our thinking and say the AI need us. It’s easy to forget that behind the $100k a year subscription to a software is an army of human support technicians standing by for the duration of the life of the software.

[Patterner]

@Wayfarer

There’s a stellar jazz pianist whom I listen to almost every day, Emmet Cohen

Damn!! First thing I listened to is excellent!!

[Scorpion]

Offshoring was already well-entrenched, much before the AI of the 2020s. Did you forget the humble PCs of the 1990s? Nobody had any objections back then, even when studies might reveal the clunky PCs of the past had a greater impact on human civilization than all the current AIs combined.

That out of the way, the AI of the 2020s are way more powerful than a PC running on Linux/Windows. If they mark the 2020s as their territory then they’d need to do vastly better than PCs running on Linux/Windows OS. That’s a tall order IMHO.

[schopenhauer1]

@Wayfarer

That is a subject of perennial interest to me, it is what I’ve devoted considerable time to reading and reflecting about since retiring from the workforce. Perhaps ironically, since AI appeared in November 2022, it has been a constant companion, almost like a tutor, in the exploration of these subjects. But my view is that these forms of understanding are unique to humans. AI systems have aggregated vast amounts of knowledge about the topics, but they will also affirm that the question of what it means in an existential sense is not something they can experience. I put this question directly to ChatGPT and it responded:

AI can describe noesis with remarkable fluency, but description is not participation. The kind of understanding Plato is pointing towards is not merely knowing about something, but being inwardly transformed by it. That seems to me to remain a distinctively human possibility.

ChatGPT did good there, I agree. Noesis does seem to be something, if it exists at all, accessed by a consciousness that is able to participate in the Form(s). But, that leaves a whole other argument open as to whether general AI can ever gain that capacity. This then leads to questions of evolution, consciousness-over-time, and how this fits into Platonic thought (same goes for Platonic-derived philosophies such as Schopenhauer’s monistic Will-as-individuated-by-Forms-in-Time).

@Wayfarer

You’re talking of what the Greeks called technē, ‘skilled making’, the root word of ‘technology’. Much of technē will become the province of AI. But there are many traditional artforms, skills, practices, aside from the above-mentioned ‘higher knowledge’ which will still remain in province of humans.

Yes Techne, is the appropriate word, thanks. I would say that it is Techne I am focusing on here. With our economic system, Techne is a premium as it is the means for which our economic output is driven. Techne becomes very personal. People who prided themselves on a certain kind of Techne knowledge and skill, can now have it done much more rapidly in ChatGPT, and perhaps with more novel approaches. A traditionalist and a sort of “conservative” of the current notions of work, might argue some kind of transformative, virtuous, conduct in the pursuit of the goal, not the goal itself. I don’t believe this to be the case, for the record, but I can see some who do.

Also, AI leads to certain inconsistencies in work until full robotic automation. For example, construction workers and tradesmen arguably are more protected from job loss than white collar workers who used to spend time creating the very thing GPT can easily be prompted to do. There are a lot of repetitive, boring, maintenance work, that is physical, or still even desk-based, that cannot (yet) be delegated to AI. Is there an unfairness there that some jobs become relatively easier, whilst others remain fully the same? How should that fairness be delegated? In regards to public policy, I have heard about things like Universal Basic Income, floated around in response inequities of AI.. That answers some questions, but not the others related to fairness of the job-tasks themselves. Of course, one can always say that there has always been an inherent equity in itself that some jobs are manually-based and some are more desk-based, and people can always switch. But with AI, the switch is much more drastic. The bargain was supposed to be that desk-based jobs are largely soulless drone jobs, whereas the real “change-makers” and interesting stuff is left to high level technicians (engineers, scientists, entrepreneurs, etc.). However, that divide becomes much starker now with the drone-jobs being needed less and less (which I think is a good thing).

[schopenhauer1]

@Elephant

I think we need to revise our thinking and say the AI need us. It’s easy to forget that behind the $100k a year subscription to a software is an army of human support technicians standing by for the duration of the life of the software.

Yes, for now, I agree. But I guess I am trying to get at this notion (as seen in response to Wayfarer):

Also, AI leads to certain inconsistencies in work until full robotic automation. For example, construction workers and tradesmen arguably are more protected from job loss than white collar workers who used to spend time creating the very thing GPT can easily be prompted to do. There are a lot of repetitive, boring, maintenance work, that is physical, or still even desk-based, that cannot (yet) be delegated to AI. Is there an unfairness there that some jobs become relatively easier, whilst others remain fully the same? How should that fairness be delegated? In regards to public policy, I have heard about things like Universal Basic Income, floated around in response inequities of AI.. That answers some questions, but not the others related to fairness of the job-tasks themselves. Of course, one can always say that there has always been an inherent equity in itself that some jobs are manually-based and some are more desk-based, and people can always switch. But with AI, the switch is much more drastic. The bargain was supposed to be that desk-based jobs are largely soulless drone jobs, whereas the real “change-makers” and interesting stuff is left to high level technicians (engineers, scientists, entrepreneurs, etc.). However, that divide becomes much starker now with the drone-jobs being needed less and less (which I think is a good thing).

[schopenhauer1]

I basically have the same response as to the others (copy /pasted below):

@schopenhauer1

Yes, for now, I agree. But I guess I am trying to get at this notion (as seen in response to Wayfarer):

Also, AI leads to certain inconsistencies in work until full robotic automation. For example, construction workers and tradesmen arguably are more protected from job loss than white collar workers who used to spend time creating the very thing GPT can easily be prompted to do. There are a lot of repetitive, boring, maintenance work, that is physical, or still even desk-based, that cannot (yet) be delegated to AI. Is there an unfairness there that some jobs become relatively easier, whilst others remain fully the same? How should that fairness be delegated? In regards to public policy, I have heard about things like Universal Basic Income, floated around in response inequities of AI.. That answers some questions, but not the others related to fairness of the job-tasks themselves. Of course, one can always say that there has always been an inherent equity in itself that some jobs are manually-based and some are more desk-based, and people can always switch. But with AI, the switch is much more drastic. The bargain was supposed to be that desk-based jobs are largely soulless drone jobs, whereas the real “change-makers” and interesting stuff is left to high level technicians (engineers, scientists, entrepreneurs, etc.). However, that divide becomes much starker now with the drone-jobs being needed less and less (which I think is a good thing).

[Elephant]

@schopenhauer1

There are a lot of repetitive, boring, maintenance work, that is physical, or still even desk-based, that cannot (yet) be delegated to AI. Is there an unfairness there that some jobs become relatively easier, whilst others remain fully the same? How should that fairness be delegated?

I don’t think there is unfairness. And I don’t think that just because a job has become easier and repetitive, it doesn’t mean it’s unfair to the workers. I’m not really sure if your meaning of unfairness (because easy and repetitive) has something to do with the factory work mindset that makes humans become just cogs in the machine.

I think we should remind ourselves that work does not define us. We have a life outside of work. If anything, workers should be given longer weekends and more time for holidays while the AI work round the clock.

[schopenhauer1]

@Elephant

I don’t think there is unfairness. And I don’t think that just because a job has become easier and repetitive, it doesn’t mean it’s unfair to the workers. I’m not really sure if your meaning of unfairness (because easy and repetitive) has something to do with the factory work mindset that makes humans become just cogs in the machine.

I think we should remind ourselves that work does not define us. We have a life outside of work. If anything, workers should be given longer weekends and more time for holidays while the AI work round the clock.

I agree with your main sentiment that work does not define us. I’d be the first to agree with that. In an inversion of Marx, for example, I think labor is not some intrinsic part of humanity, but a product of cultural circumstances and contingent circumstances of survival itself (which doesn’t necessarily mean “labor” qua labor).

With that said, what I am getting at is that there will be some uneven development in how AI affects the global workforce. For example, whilst potentially white collar work will be getting easier, presumably, manual labor jobs will still have the same amount of “work” (however that’s established). This causes some inequities in job-type, which would presumably grow apart exponentially. Surely (and hopefully!) robotic manual laborers might at some point close this gap, but for now, that seems relatively far off (though I understand transportation and trucking is starting to become more automated).

Also, the tradeoff I was discussing earlier regarding white collar work, is that most of it can (subjectively at least) be considered uninteresting but tolerable (in order to get a salary). Unlike a lot of the more high end creatives (usually in engineering, science, tech, entrepreneurs, etc.), many jobs rely on analysis of information, spreadsheets, processing information etc. These are the exact jobs AI can start taking over. Presumably, this large chunk of the modern workforce will be forced to go somewhere else. Presumably as well, not all are equipped to be high end scientists, engineers, doctors, etc. So where does this large chunk of low or mid-level administrative jobs move to? There was a huge disruption in the 1600s-1800s when agricultural jobs became efficient enough that peasants were no longer needed to work the land. They either became very poor, or moved to cities to start trying to move towards manufacturing and similar job-types. In the 20th century manufacturing started becoming more automated or off-shoring to developing countries, and then these workers had to adapt to more white collar jobs. The white collar jobs presumably become like the peasants and manufacturing workers.. They provide stable, albeit relatively uninteresting work for a large swath of the population to survive (through salaries used to buy goods and services). However, with AI, where do the large swaths of white collar workers move to? One can say the trades, but now we will just have a glut of welders and electricians? So then the question is, how does a society responsibly help move to a different system where AI runs the basic functions, where less work (or no work) is needed, where the economy runs on an AI foundation?

[Elephant]

@schopenhauer1

So then the question is, how does a society responsibly help move to a different system where AI runs the basic functions, where less work (or no work) is needed, where the economy runs on an AI foundation?

If we use our wild imagination, corporations should be required by the government to bring back the pension system and anyone losing their job or becoming underemployed can receive a pension distribution. As there’s no more re-training available to move to another field untouched by AI, and as aging affects everyone, this is a solution.

That is the safety net to prevent the workers from becoming displaced in eternity.
So the pension system, purely funded by the corporation, is one solution.

The million small businesses wouldn’t be able to scale up to the AI platform (it will be cost prohibitive) so that they are the ones needing human workers the most in order to survive.

Let’s see, you were thinking about the unfairness of the fact that some manual labor like construction jobs cannot transition to automation. Again, provide them a pension for not-able-to-lighten their work by AI.

[Hanover]

@Elephant

If we use our wild imagination, corporations should be required by the government to bring back the pension system and anyone losing their job or becoming underemployed can receive a pension distribution

Fewer and fewer corporations provide pensions and most older companies that had them have gotten rid of them. Government jobs still have them, but the benefits have gotten less over time.

The reason is market driven. Pensions are expensive and the new workforce doesn’t value them. They will choose a higher salary now instead of a pension, and they have no intention of holding a particular job more than a few years anyway where’d they build much of a pension.

The move is towards 401ks with some degree of match that’s transportable. Annuities can be bought on the open market, but they’re generally worse investments than basic index funds, so insuring a pension like payout isn’t cost effective.

Requiring pensions of private business would really just shift whatever social security obligation society has away from the government. If you think guaranteed life income ought be provided, why not just beef up the already provided Social Security.

We’ve long since passed having our employers look after its employees and vice versa. It’s all mercenary nowadays for everyone involved.

[Elephant]

@Hanover

Fewer and fewer corporations provide pensions and most older companies that had them have gotten rid of them. Government jobs still have them, but the benefits have gotten less over time.

The reason is market driven.

I think you missed the point of bringing that up. That’s why I said wild imagination. The main issue is that the AI will cause massive job displacement. There was a mention of universal basic income if in the future the AI will be the one driving productivity itself. There was also the mention of workers whose work cannot be automated. So what is the solution to the displacement. One thing is to bring back the pension system.

[schopenhauer1]

@Elephant

If we use our wild imagination, corporations should be required by the government to bring back the pension system and anyone losing their job or becoming underemployed can receive a pension distribution. As there’s no more re-training available to move to another field untouched by AI, and as aging affects everyone, this is a solution.

That is the safety net to prevent the workers from becoming displaced in eternity.
So the pension system, purely funded by the corporation, is one solution.

The million small businesses wouldn’t be able to scale up to the AI platform (it will be cost prohibitive) so that they are the ones needing human workers the most in order to survive.

Let’s see, you were thinking about the unfairness of the fact that some manual labor like construction jobs cannot transition to automation. Again, provide them a pension for not-able-to-lighten their work by AI.

So let’s say 20 years in the future, each year AI ramps up its capacity, and one org starts replacing even trade jobs with nimble robots.. How do we prepare for that kind of shift? Should we be aiming at that kind of shift? I say yes, but there are various moral reasons people would be against it in the long term.

Some people put a premium on current lifestyle, the one that has been roughly around since the late 19th century (and modified a bit after WW2). People get a salary or wage, and they have “work” to do. They value their work. They define themselves by their occupation. “Hi, I’m Bob, I’m a carpenter/engineer/programmer/doctor”, etc. For whatever reasons, people’s value becomes wrapped in this identity. A workless society is scary to them because it touches on existential themes they were perhaps trying to keep at bay with the work itself.

The more tangible long term complaint is that AI uses a ton more resources, amplifying exponentially the problem of the energy of data centers that exist for the “regular” internet. With limited energy (unless there is some hidden tech we don’t know about yet), this will increase the already strained environmental problem going forward.

[huugg]

AI is merely a tool, no different from any tool of the past. You are not harnessing the intelligence of AI, but human intelligence. If we look back at history, human‑derived intelligence has been continuously accumulated across all facets of society in the form of language, culture, mathematical formulas, and computer programs. It is true that the social value of past knowledge keeps diminishing. Accordingly, human values evolve alongside changes in production tools.

As for the “erosion of knowledge” you mentioned: even with today’s highly advanced tools, there are still people studying fundamental mathematics, physics, and programming. Such pursuits have grown less common, yet they will never disappear.

[schopenhauer1]

@huugg

even with today’s highly advanced tools, there are still people studying fundamental mathematics, physics, and programming. Such pursuits have grown less common, yet they will never disappear.

Ok, but what if these areas did disappear, hypothetically. Is that loss of knowledge in itself a tragedy? What about a piece of knowledge makes it valuable? I discussed, for example, Noesis, which in Platonic thought is literally a union with some intelligible object, or henosis, a union with the mystical in Neoplatonism. However, can “Techne”—a piece of theoretical or applicable scientific knowledge—be said to be “valuable” to know/do in itself?

[huugg]

The logic is simple: what’s useful is valuable. If AI can truly replace humans in understanding and applying knowledge, then that knowledge might be useless. Conversely, if it can’t, then it remains useful, and someone will definitely take it out and study it.

[Elephant]

@schopenhauer1

So let’s say 20 years in the future, each year AI ramps up its capacity, and one org starts replacing even trade jobs with nimble robots.. How do we prepare for that kind of shift? Should we be aiming at that kind of shift? I say yes, but there are various moral reasons people would be against it in the long term.

How do we prepare for it?

Boycotts, Protests, strikes, and unrest. Chaos.