The AI Daily Brief: Artificial Intelligence News and Analysis - Why We're Radically Underestimating the Political Implications of AI Job Displacement

Episode Date: August 6, 2023

NLW opines on the increasingly popular canard "AI won't take your job; someone using AI will" and finds it lacking. ABOUT THE AI BREAKDOWN The AI Breakdown helps you understand the most important news and discussions in AI.  Subscribe to The AI Breakdown newsletter: https://theaibreakdown.beehiiv.com/subscribe Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@TheAIBreakdown Join the community: bit.ly/aibreakdown Learn more: http://breakdown.network/ Twitter: https://twitter.com/nlw / https://twitter.com/AIBreakdownPod

Today on the AI breakdown, we're discussing why we're not taking seriously enough the likelihood of political upheaval that could come from AI-driven white-collar job loss. The AI breakdown is a daily podcast and video about the most important news and discussions in AI. Go to Breakdown.network for more information about our Discord, our YouTube, and our newsletter. Welcome back to the AI breakdown. Today we're talking about a thesis that I have, and that is that we are radically underestimating the likely political upheaval. from AI job disruption. Now, on the one hand, this might seem a little bit surprising. There are articles seemingly every day, certainly every week about the impact of AI on jobs. It's something that comes up every time there's a congressional or Senate hearing about the topic. And yet, even with that,

I think that one, at least for the last few months, the extinction risk question of AI has taken front and center relative to the job discussion. And two, when it comes to the job discussion, the discourse is just really, really underwhelming. So why is that? I think that part of it is that we don't really have any precedent for a technology disruption or automation that impacts not lower class or middle class workers, but actually cuts directly at upper class workers and nullifies many of their skills. Now, what we do have recently is an example of what happens when you gut the middle class.

In this case, not necessarily as a byproduct of automation, but instead as a byproduct of automation, but instead as a byproduct of globalization, and what it led to was the most disruptive political movement of the last 30 years. Part of the reason that I think we're underestimating the political upheaval to come is that the cohort of people that it affects are structurally politically more powerful and likely to fight a hell of a lot harder with more resources to do so. Let's talk about the state of the discourse right now. I think there are two problems, broadly speaking.

The first is a tendency to zoom too far out and gloss over what happens in the meantime. I think a great example of this is Sam Altman's testimony at the first Senate hearing on AI that happened a couple of months ago. He said, I expect there to be significant impact on jobs, but exactly what it looks like is very difficult to predict. As our quality of life raises, and as machines and tools that we create can help us live better lives, the bar raises for what we do. I'm very optimistic about how great the jobs of the future will be. He went on and said, GPT4 and other systems like it are good at doing tasks, not jobs. And so you already see that people using GPT4 to do their job much more efficiently by helping them with tasks. Now GPT4 will, I think,

entirely automate some jobs, and it will create new ones that we believe will be much better. Mark Andreessen also addressed this in his piece why AI will save the world. He wrote, the fear of job loss variously to mechanization, automation, computerization, or AI has been a recurring panic for hundreds of years, since the original onset of machinery such as the mechanical loom. Even though every major new technology has led to more jobs at higher wages throughout history, each wave of this panic is accompanied by claims that this time is different. This is the time it will finally happen. This is the technology that will finally deliver the hammer blow to human labor.

And yet, it never happens. We've been through two such technology-driven unemployment panic cycles in our recent past, the outsourcing panic of the 2000s, and the automation panic of the 2010s. Notwithstanding many talking heads, pundits, and even tech industry executives pounding the table throughout both decades that mass unemployment was near. By late 2019, right before the onset of COVID, the world had more jobs at higher wages than ever in history. Now, where I agree with Sam Altman and Mark Andreessen is around the ultimate, the end state. I agree with Andreson when he writes that, quote, The core mistake the automation kills jobs, doomers keep making is called the lump of labor fallacy.

The fallacy is the incorrect notion that there is a fixed amount of labor, to be done in the economy at any given time, and either machines do it or people do it, and if machines do it, there will be no work for people to do. Instead, Andresen writes, technology empowers people to be more productive. This causes the prices for existing goods and services to fall and for wages to rise. This in turn causes economic growth and job growth while motivating the creation of new jobs and new industries. If a market economy is allowed to function normally, and if technology is allowed to be introduced freely, this is a perpetual upward cycle that never ends. Now, Sam's way better jobs in the future is basically saying the same thing.

I agree wholeheartedly that economies aren't some basal set of X amount of activity and that if that activity takes less labor, there's going to be nothing for people to do. Markets will always fill in gaps in ways that are totally unanticipated. I think it is absolutely the case that entire categories of jobs that don't exist today will employ millions in like a decade. Where I disagree, however, and where I find the discourse lacking is in the transition. Remember, the thesis of my piece, is that we're underestimating the political upheaval from AI job disruption. It is not that AI will take all our jobs. It's that the process of broad economic transition wrought by AI

is going to have massive implications for entire industry's careers and professions in ways that will inevitably lead in the transition period to massive societal and political battles. Ghibly put, creative destruction still involves destruction. I also find the least compelling part of Andresen's argument that the real conclusion of the, quote, outsourcing panic and automation panic was that the world had more jobs at higher wages

than ever in history just before COVID. Ironically, given Mark's later arguments, this completely reduces people's employment to just a lump of labor once again. The idea that there has been no meaningful impact because of outsourcing is ludicrous. It would take a willful denial of watching the last decade of politics

to actually make such an argument. Now, to the extent that one believes that society's only obligation to its people is that there is some type of job, for everyone, then fine. Maybe we decide to measure things entirely on the basis of just employment in general. But between the Tea Party, Occupy Wall Street, Bernie Sanders, Bitcoin, the idea that the mass transformation of the nature of the American economy and the global economy over the last 20 years hasn't had an impact is just nuts. And yet I digress, because frankly, Sam Altman and Mark Andresen

are still much better in terms of the discourse around this topic than the LinkedIn pundits who are saying the single stupidest thing associated with this entire conversation. Can you guess what I'm thinking of? That's right. It's AI won't take your job. It's somebody using AI that will take your job. Do me a favor. Just go do a Google search for that phrase. See how many silly op-eds and LinkedIn think-boy pieces use that exact canard to try to sound clever. So I have two big complaints with this. The first is slightly pettier, I will admit. And it is around who I'm are the main progenitors of this statement. TLDR, it's the boomer social media hacks that are basically the professional equivalent on LinkedIn of people who put inspirational quotes over nice pictures on

Instagram. It's saccharin. It's designed for engagement. And it's predicated on a social media business model that's basically just about sounding clever. Now, the second reason that I find this phrase so detestable is that it's completely not true. Yes, it is the case that almost every job will be AI enabled, so yeah, if you can't use AI, you'll be lacking a major skill set. Someone else that does what you do that also knows how to use generative AI tools is likely going to take your jobs, I will admit. But this glib response completely denies the reality that entire categories of things that people do now are just straight up going to be done by machines. For example, let's talk about basically everything around performance advertising on major social media

platforms. All of it, from targeting to creative, to resource allocation and budgeting, every aspect end to end of performance marketing, representing hundreds of billions of dollars a year in advertising spend, and hundreds of thousands of jobs, is being pushed every single day away from people and towards AI on the big platforms. That means that very, very practically, thousands and thousands of people who are today specialists in Facebook ads are not just threatened by other Facebook ad experts that are good at using chat GPT for copy, they're threatened by the platform AI itself. This whole feel-good sound clip fortune cookie wisdom of it's not AI that's going to take your job it's people using AI is at best temporary self-delusion to allow us to catch up with the reality

of so much impending change. And at worst, just ostrich head in the sand denialism. Want some additional evidence beyond just my opinion and one job category example? Let's look at a recent study by KPMG, Gardner, Fortune, and Goldman Sachs research. they asked Fortune 500 CEOs how they expected AI to impact company headcount. Now, it's true when the question was framed in terms of just next year, over 40% said that they thought that the labor need would be unchanged, right? That wasn't going to really have a big impact. That compares to just around 20% who said that they thought that perhaps lower labor would

be needed because of AI. But zoom out five years. The percentage of Fortune 500 CEOs that said that labor needs will be unchanged because of AI dips to well under 20% while the percentage of Fortune 500 CEOs who believe that there will be less labor needed increases to more than 70%.

70 plus percent of Fortune 500 CEOs anticipate that AI will lead to lower labor needed. You'll notice that that stat doesn't say that more than 70% of Fortune 500 CEOs think that they will have the same headcount but that people who use AI will replace people who can't use AI

within their organizations. Now, just to really scare the crap out of any American professional still listening at this point. All of this AI stuff is converging with another trend, which is the fact that the increase in acceptability of remote work and work from home is also having companies think differently about how they view global talent as opposed to expensive US talent. Put simply, if you create processes that make it such that you don't need someone in an office to do a job, is there someone who can do that job just as well with Native English equivalence for a quarter of the price somewhere else in the world? Now, of course, just like with AI,

it's too early to know exactly how those trends play out. But the point is, you have multiple sources of pressure for this upper wage cohort, in again, ways that we just haven't seen yet manifest. And so if and when we see big categories of white collar and upper wage jobs threatened by AI, what might be some of the likely political responses? Well, I certainly think that right now we're getting a preview of one of them, which is the resurgence of labor organization. Now, yes, it is of course true that the writer's strike and the screen actors guild strike

is not just about AI. But in the context of that strike, AI is the embodiment of what those unions view as a fundamental disrespect and commoditization of their work coming from the entertainment studios. Now, it may be a tactic for those unions to basically reach out to the rest of the world and say this fight is coming for you next, but I also don't think it's exactly wrong. It seems very, very likely to me that one of the major political shifts that will happen as AI starts to infringe on more of these white-collar upper-raged jobs is a resurgence in labor organization. Attempting to collectively make it so that management has to pay a much greater cost for a lower-cost technology is in many cases going to feel like the only option for many of these different industries.

Going back to the thesis of this show, even if that new labor movement is successful, we're at a historic low in labor organization. Shifting back into a unionized world is not going to be clean or easy. In other words, it's going to involve a lot of political upheaval. Now, speaking of not clean and easy, what even are those unions really going to demand? That governments say to private enterprise, no sorry you can't take advantage of new technology, that the U.S. Department of Commerce makes it illegal for shipping companies to use self-driving trucks instead of manned vehicles? Where do they draw those lines? And holding a sign how difficult it is to draw those lines, industry by industry, enforcing them will almost definitionally mean a significant increase in government power.

And by the way, that's not the only way that AI is already nudging state power higher as well. First of all, some of the AI safety folks are outright fine with autocracy as long as it's human. Andresen again tweeted in March, If AI really is existentially dangerous, then it can't be worked on anywhere as the risk of a lab leak is still catastrophic. The only answer is a global authoritarian dictatorship of cosmic proportions. Now, he is, of course, trying to make a point that that's bad, but I've seen plenty of folks on the opposite side of that coin, ready to sign away their rights to autotropic. as long as it's an autocracy controlled by humans.

And even if we get away from individuals, extinction risk is of course exactly the type of political issue that can lead to people handing over more power to the state than they might otherwise would. Now, other folks aren't just going to want to cede that power to the state willy-nilly. So once again, here we are set up for a significant political battle. So what are ways to deal with this? If we accept this thesis,

that even if things are great ultimately, the transition to an AI-enabled world is going to cause massive political upheaval that is the fallout of massive economic upheaval, how do we get out ahead of it? One is, of course, a massive focus on the transition. We're talking a Marshall Plan for Reskilling, education, training. This is the type of thing that's politically popular to talk about, but doesn't really go anywhere. But if UBI made it to the last presidential campaign as a discussion point of the potential fallout of technology automation, maybe actually trying to retrain the entire workforce to be part of the new economy is something that someone might want to take on as an

actual viable issue. Now, second, it strikes me that there may be some political alignment here between the AI safety folks who think we should pause and the people who are concerned about this sort of economic fallout. Basically, even if we did nothing from here, we didn't advance at all past GPT4, big swaths of jobs and roles and professions would be totally changed. However, obviously, it's the things coming down the pipeline that have people really concerned. There might be an AI safety AI jobs disruption axis that basically agree on having everything beyond this point much more tightly controlled. The argument for this would be that doing it preemptively, creating a gauntlet that AI companies and models have to go through, controlling the speed with which it's rolled out

and made commercially and widely available, would be easier to do with less state-level power creep than the sort of draconian response it would take to try to put that toothpaste back in the tube once it's squeezed. I don't know if I'm ready to sign on for this yet, but it's at least something that's interesting to discuss. Ultimately, I will fully admit that I don't know. I don't know which jobs are going to go away. I don't know what the impact will be. I don't know how long or how painful the transition to a better economy will be. But I do feel confident that there's going to be a transition and one that inevitably involves some serious pain. And on top of that, what I also do know is that we're not giving it the thoughtfulness it deserves. I think we have a lot

of work ahead of us. And if no one ever again said AI won't take your job, but someone using AI will, it would be too soon. Anyways, guys, rant over. Let me know what you think. Come join us in the AI Breakdown Discord. Links down there in the show notes or go to Breakdown.network. I appreciate you listening or watching as always. And until next time, peace.