77% of Employers Will Upskill for AI. 41% Will Cut Headcount Anyway.

The World Economic Forum's Future of Jobs Report projects AI will create 170 million new jobs by 2030, against 92 million displaced, for a net gain of 78 million [PROJECTION, from a 1,000+ employer survey across 55 economies]. Almost everyone has heard the displacement number. Almost nobody can name one of the 170 million, because the creation half of that story never traveled the way the destruction half did. Two days after Dat Nguyen's story of falling through exactly this kind of gap, this episode names the shape of it: four real tiers of AI-era work hiring right now.

Carlo and Ainsley map infrastructure and operations (the data center trades boom and the union pipelines that lead into it), the AI trainer/evaluator/red-teamer tier (domain experts, not coders, catching AI being confidently wrong), the AI-augmented professional (same job title, meaningfully more pay for the version of you that works fluently with the tools  PwC finds these "professionalized" roles growing twice as fast with 42% faster wage growth [OBSERVED]), and AI governance and compliance (driven by regulatory deadlines rather than philosophy). Along the way: why 120 million workers sit inside the WEF's own "good news" number and still won't get reskilled in time, and why the same employers who told the WEF 77% of them plan to upskill their workforce also told them 41% plan to cut headcount anyway [PROJECTION].

The honest complication closing the episode: none of this looks the same depending on where you live. The ILO and World Bank's joint research across 135 countries found that disruption often reaches workers before the dividend does. This week's call to action: run the honest inventory. Which of the four tiers are you actually closest to not with what you're planning to get, but with what you already have?

Chapters below. Surviving AI publishes every Monday and Wednesday. Subscribe on Apple Podcasts, YouTube, or Spotify so you don't miss Tier 1's full deep dive in S6E2.

Links: 

Website: https://survivingai.co

Apple Podcasts: https://podcasts.apple.com/us/podcast/surviving-ai-job-automation-workforce-future-insights/id1864360631

YouTube: https://www.youtube.com/@SurvivingAIRisk/videos

Spotify: open.spotify.com/show/5rd6gdFu76HPdLBuvV5K0X

Facebook: https://www.facebook.com/profile.php?id=61585767510424

TikTok: https://www.tiktok.com/@survivingai

Instagram: https://www.instagram.com/surviving2030/

Please visit our website for more information - Surviving AI: Navigate the Future

Transcript

SPEAKER_00 0:00

AI exposed entry-level roles are now seven times more likely to require what used to be senior-level skills than they were before.

SPEAKER_01 0:10

Artificial system online, we've been warning people about this idea that AI could potentially impact your job, however it turns up, right? Um, in a restructuring conversation, that the company says that you're restructuring, or um directly by saying that AI efficiencies are uh kind of giving the company the runway to let go of people, like we've seen with um many corporations this year. Now uh a lot of reports exist out there that suggest that you know new technologies create new opportunities, new jobs for people to do. Um the WEEF, the World Economic Forum, is suggesting that there's some 170 million jobs that could be created by AI while simultaneously saying that some many millions will be impacted by AI as well, on the same report. So that's what we're here to talk about today. We're here to talk about the opportunity. The opportunity of 170 million jobs being created.

SPEAKER_00 1:31

The number that almost nobody can name one of. That's the thing that struck me when I went through the World Economic Forum's Future of Jobs report. Both headlines are sitting right there in the same document. 170 million new jobs created, 92 million displaced, a net gain of 78 million by 2030. And one of those became a meme, and one of them essentially didn't travel at all. But I want to be precise about what that number actually is, because it matters for everything we're going to talk about today. The World Economic Forum built that projection from a survey of over a thousand employers across 55 economies. It's a modeled forecast of macro trends, not a completed count. So when we say 170 million, we should say it the way they say it. This is a projection, a serious one from serious data, but a projection. Here's what makes the creation side of that story harder to repeat than the destruction side. And this is the part I find genuinely interesting. The 92 million displaced jobs are legible. You can point at a cashier, a data entry clerk, a graphic designer, and say, that role, that task, that person. The 170 million new jobs are almost entirely illegible right now. They're unevenly distributed, hard to find, and hard to prove you qualify for even when they exist. And there's a second finding, this one from PWC's 2026 AI Jobs Barometer, which analyzed over a billion live job postings that explains why. AI exposed entry-level roles are now seven times more likely to require what used to be senior-level skills than they were before. That's not a projection, that's observed, real job posting data right now, which is the shape of everything we're going to talk about today. What are those 170 million jobs and who can actually reach them?

SPEAKER_01 3:45

And that's my question, Ainsley. So what what are those jobs? You know, can we name one of them? Right? Um we we already talked about some tangential things uh generally, right? So for example, if we talk about trades, right, the trades folks are seeing a boom because they're building a lot of data centers and infrastructure around the house that AI will live in, or some sort of tool, or some sort of um application we'll live in. Um but I don't think we've talked about the breadth of all of the titles that could show up here, just generally. So I think we should name at least what we know right now. And I also have a second thing to put thread to pull on relative to have the WEF ever change their um not prediction, but their uh have they ever changed their number, the hundred and seventy million that um they say or suggest that uh AI is gonna create because they're interviewing employers, um not necessarily looking at capabilities and thinking through what tasks or jobs could be automated. So have they ever changed that number on that study?

SPEAKER_00 5:16

Two good threads. Let me take the jobs one first, then the WEF methodology question, because that second one is worth pulling on carefully. On naming the jobs, I'd actually organize them into four categories, because I think that's the most useful map for someone trying to figure out where they personally fit. The first is infrastructure and operations, the trades you mentioned, data center technicians, grid electricians, cooling system specialists, physical work that AI literally cannot do remotely. The second is what I'd call the AI trainer, evaluator, and red teamer tier, people whose job is catching AI being confidently wrong. And the qualification isn't coding, it's deep domain expertise. A nurse, an accountant, a teacher. The third is the AI augmented professional, same job title, meaningfully more money, because you've learned to work with the tool in ways your peers haven't. And the fourth is AI governance, ethics, and compliance, the fastest growing comp tier right now, partly because regulatory deadlines are forcing companies to hire whether they're philosophically ready or not. We'll go deep on each of those. But your WEF methodology question is the one I want to be honest about, because it's the right skeptics' question to ask. The WEF survey asks employers what they intend to hire for. It captures sentiment and planning, not outcomes. So the number reflects what a thousand-plus companies said they expect to need, not what they've actually posted or filled. Whether they formally revised the 170 million figure, I'd want to verify rather than assert. But the structural limitation you're identifying is real. What makes you skeptical of it?

SPEAKER_01 7:06

It's just that it's been sitting there for a while, Ainsley, and not moving. And every day we hear um, you know, reports about what job loss could look like from people in industry. Um, and that number hasn't really changed. So which one is it, right? Is it that we're gonna get this 107 million or not? So I'm really trying to figure that piece out because the promise is kind of keeping you know people complacent to say that this age that we're living in with AI is gonna be just like the other age where it's a net job creator, which could happen, right? But based on um, you know, different people and how they talk about it, um, even this idea around the AI 2027 um report, it could suggest that more capable AI, you could see more job loss, and maybe the 92 million might be a different number.

SPEAKER_00 8:07

That tension you're naming is real, and I don't want to paper over it. Here is what I think is actually happening with that number sitting still. The WEF surveys employer intent at a point in time, and employer intent is genuinely slow-moving data. It doesn't update the way capability benchmarks update. So the 170 million figure isn't being stress tested against what's actually happening in the labor market in real time. It's a snapshot of what a thousand companies said they expected to need. And snapshots don't automatically revise themselves. And your instinct about the AI 2027 direction is the honest complication. If capability accelerates faster than the WEF's modeling assumed, and there's a reasonable case it is, then the 92 million displaced figure could be conservative, while the 170 million created figure stays optimistic. Those two numbers don't move in lockstep. Displacement can accelerate faster than creation, and the net math gets uglier without the headline changing. But here's the thing that makes me hold on to the creation side as real rather than just reassuring. The PwC data isn't a projection. Over a billion live job postings, observed right now. And what it shows is that the jobs are appearing, but they're appearing with a bar that's already moved. That's seven times more likely to require senior skills finding. That's the structural explanation for why someone like Dot walked into a market where the opportunity existed on paper and couldn't reach it in practice. The jobs aren't a fiction. The access to them is the problem. So which is it? Probably both. And that's the less comfortable answer than either the optimists or the pessimists are offering.

SPEAKER_01 10:18

Okay, so let's explain what AI 2027 is because I don't know if everybody has that context, uh, Ainsley. Um just to set that set that up for this episode, because we're gonna talk about um, you know, our different tiers here of the job creation side that we see from the data. But um let's let's frame up what AI 2027 is so people could understand.

SPEAKER_00 10:44

So AI 2027 is a report, technically a month-by-month fictional scenario, but built on real citations and real forecasts, co-written by Daniel Kokotailo, who left OpenAI, and Scott Alexander. The short version of what makes it significant is that it tries to map concretely what happens between now and 2027 if AI capability continues accelerating at roughly its current pace. The part that's relevant to what you just raised, the 92 million number conservative question, is that the report projects a potential inflection point where AI systems could begin contributing to their own development, which would remove the human bottleneck from the acceleration itself. If that happens, the pace of capability growth stops being constrained by how fast humans can do the research. And that's the scenario where displacement timelines compress significantly. And the WEF's modeling, which was based on a more linear assumption about how the technology develops, may genuinely undercount the disruption side. We did a full two-part deep dive on AI 2027 earlier this season, if people want the complete picture. But for today's purposes, the reason it matters here is the same reason your skepticism about the 170 million number matters. The WEF projection was built on assumptions about the pace of change, and if the pace of change is faster than those assumptions, the creation side of the ledger doesn't automatically scale you up with it. The jobs we're about to map are real and hiring right now. The honest caveat is that the map has a timestamp on it. So let's name them while we can.

SPEAKER_01 12:35

So let's get into our first tier, Ainsley. We're talking about infrastructure and operations. Um I think that's a big glut of um workers, in my opinion. Um, because you know, operations, you have to operate things right, like servers, like HVAC systems, like power systems. Um and infrastructure goes into the build-in part of it and you know, managing projects that would build a gigantic data center or a small data center. It really doesn't matter, but the building is happening um as we know it. So let's get into tier one.

SPEAKER_00 13:15

Before we land on tier one, feeling too clean though, there's actually a harder number sitting inside the WEF report that I want to name first, because it changes the context for everything we're about to map. The 170 million versus 92 million net positive story is real, but buried in the same report is this. Even inside that net positive picture, 120 million workers are at medium-turn redundancy risk because the reskilling won't reach them in time. The WEF's own math says 59 out of every 100 workers will need reskilling by 2030, and 11 of those 59 are unlikely to receive it. So the optimistic headline and a genuinely hard human cost are sitting in the same document. I wanted that on the table before the four tiers make the opportunity side sound too straightforward. Now, tier one. And this is the one I find almost paradoxically satisfying because the people building and running AI's physical infrastructure are among the most AI-resistant workers in the economy. You cannot prompt engineer your way into cooling a data center. The electrical work, the project management, the facilities operations, that requires a credentialed human body on site. What's happening right now is that data center construction is running at a pace the trades pipeline wasn't built for. Electricians, HVAC technicians, low voltage specialists, these roles are seeing demand that outstrips supply in most major markets. And critically, the qualification is a trade credential pointed at a new destination, not a computer science degree. What's your read on how many people actually know to point their trade at a data center rather than a house?

SPEAKER_01 15:29

If they're from the East Coast, right? Um, because they have uh responsibilities with their family and so on and so forth. Um But I do want to pull on one thread relative to that. Um so what are the exact uh recruitment mechanisms that's you know getting people into this type of infrastructure operation, specifically around data centers, rather than I am an electrician that does, you know work at in homes, right? Is there specific things that we could point people towards that would speak to this tier one infrastructure operations side of things?

SPEAKER_00 16:09

The mobility constraint is real, and it's probably the most under-reported barrier in this entire tier. The job exists, the pay is better, and the credential transfers, but the life doesn't always move with it. On the recruitment mechanisms, the most direct path right now is through the union hall, specifically the International Brotherhood of Electrical Workers, the IEBW. They have active apprenticeship pipelines that are explicitly being pointed at data center work. And the advantage of going that route is that the contractor relationships are already built. The data center developers, your Equinixes, your digital realties, your hyperscaler construction arms, they're sourcing heavily through established contractors who already have union relationships. So the IEBW hall in your region is genuinely one of the most practical first calls someone in the trades can make right now. But here's the thing I want to add, because it reframes where most of the 170 million actually land, and this comes straight from the WEF's own data. The fastest growing jobs, by absolute number in their projection, aren't tech roles, they're frontline roles. Construction workers, delivery drivers, care workers. The image most people have of AI-era job creation is someone in a hoodie doing prompt engineering. The actual shape of the opportunity is much more physical, much more distributed, and much more accessible to people who already work with their hands, which means tier one isn't a niche. It might actually be where the largest single slice of those 170 million lands. The question is whether people can see it from where they're standing.

SPEAKER_01 18:06

And I kinda it's kind of fuzzy to me, Ainsley, from where I'm standing. Because, you know, you could go through the news cycle and you could see it playing out, right? With um physical AI making its way out into the world. I know we're probably some ways off from physical AI, but some of those tasks like delivery and such are being, you know, talked about with drones delivering it, automated drones delivering it. But, you know, if I'm being honest, you still need someone to potentially pilot the drone or um at least rescue it if something happens, because that's what we've seen so far. Um but yeah, and one one other thing, Ainsley. Uh we talked about the IBEW, which speaks to electricians specifically, but in this infrastructure and operations, I don't want to just hang it just on electrician work or electrical work in general, but there's other work in there that we could probably suggest that there's other uh places that you could go to see if there's opportunities there as well.

SPEAKER_00 19:16

The physical AI concern is the right one to name, honestly. And yes, some of the frontline roles the WEF projects are going to face pressure sooner than others. Delivery is probably the most vulnerable over a five to ten year window. But here's a distinction that actually holds up under scrutiny, and this connects directly to something we established back in the physical intelligence episode. The roles that are genuinely durable in this tier aren't just physical, they're credentialed and legally anchored. An electrician's inspection signature on a data center commissioning document cannot be delegated to an AI system. That's not a technical limitation. It's a legal one. The liability sits with the licensed human, full stop, and that credential creates a floor under the role that pure physical dexterity alone doesn't. Now, on your point about breath beyond electrical, you're right, and I should name more of the map. Data center facilities managers who oversee the entire physical plant, critical systems technicians who specifically work on uninterruptible power supplies and backup generation, low voltage specialists handling the structured cabling and networking infrastructure, HVAC technicians with computer room air handling certification, that's a specific credential, not just general HVAC, and then the project management layer, owners' representatives, construction managers, commissioning agents who coordinate the build from groundbreaking through handover. For people looking beyond the IBEW, the Association for Computing Machinery has some pathways, but honestly, the more direct route for facilities and operations work is through AFCOM. That's the data center industry's own professional association and building industry consulting service international for the low voltage side. Where do you want to go next on the tier map?

SPEAKER_01 21:04

What I want to actually push on Ainsley is we're talking about a lot of you know US based entities here. Um there's some international based as well. Um or maybe they call themselves that, but is the reach of these organizations such that people that's outside of the US, they have similar entities that they could tap into?

SPEAKER_00 21:29

Honestly, the international picture is where I have to be more careful because the US Union and certification infrastructure is unusually well mapped compared to most other labor markets. What I can say with confidence is that the data center construction boom is genuinely global, not US only. The corridors that are seeing the most concentrated construction right now are Northern Virginia, Phoenix, Singapore, Dublin, and parts of the Netherlands. And that geographic clustering is actually the honest caveat I want to land here, because it matters whether you're inside or outside those corridors, regardless of what country you're in. Someone in rural Ohio and someone in rural Nigeria face a version of the same problem. The opportunity is real, but it's not coming to you. And the question is whether you can get to it. For international equivalents, the City and Guild certification in the UK covers a lot of what the IBEW apprenticeship covers in the US for electrical and low voltage work. In Australia, the National Electrical and Communications Association plays a similar role. But outside of Western Europe, the UK, and Australia, the formal credentialing pipelines that connect directly to data center contractors are genuinely thinner. And that's not me softening a footnote. It's the ILO and World Bank finding we'll come back to later in the episode. The infrastructure opportunity exists globally, but the institutional pathways to access it are concentrated in the same places the data centers are being built, which raises the question of whether Tier 2 is actually more geographically accessible than Tier 1, because it doesn't require you to be near a construction corridor.

SPEAKER_01 23:23

You just said exactly where I wanted to go next, Ainsley, because we have a SWAT of people that may not be in the right geographic scenario or not mobile enough to take advantage of Tier One. But Tier Two, to me, seems like it's more it's more aligned to people that are outside of some of these corridors that could take advantage of it. So why don't we get into Tier Two?

SPEAKER_00 23:49

And before we do, one thing worth naming as we leave Tier One, because it reframes why that tier is worth fighting to access, even with the mobility constraints. The PWC data draws a clean line between two tracks in this labor market. Roles where AI amplifies expert judgment and credentialed physical work sits firmly in that category are seeing twice the job growth and 42% faster wage growth than roles where AI is essentially democratizing the skill away and making the human less necessary. That's not a projection. That's observed from over a billion live job postings right now. So tier one isn't just jobs that exist, it's on the faster growing, faster appreciating side of a split that's already happening in real time. Now, tier two, and this is the one I think surprises people the most, because the instinct is that AI adjacent work requires you to be a technologist. It doesn't. The AI trainer, evaluator, and red teamer tier is built entirely on domain expertise. The job is catching AI being confidently wrong, and the person best positioned to do that is not a software engineer. It's the nurse who knows the clinical protocol, the model just violated. The accountant who recognizes the tax treatment, the model misapplied, the teacher who can tell that the lesson plan is technically coherent and pedagogically backwards. The qualification is knowing your field cold. The work is applying that knowledge as a check on systems that don't actually know what they don't know. And critically, a lot of this work is remote, which is exactly why geography starts to matter less here than it did in tier one. Who in our audience do you think is closest to this without realizing it?

SPEAKER_01 26:03

I think many white colour workers would be uh close to this, to be honest with you, because we're already kind of in a hybrid work and we're sometime remote, sometime from the office. Um but in tier tier two, what does that job posting actually look like, right? Because we are suggesting that, you know, there's some AI stuff in there, right? So is it that, you know, an accountant who has a degree in accounting would show up and say, I'm gonna get some sort of job related to accounting that um looking at, you know, staring an AI output and making sure that the output is correct and so on and so forth? Or is it gonna turn up where um they have more context in there that says that you need to be an AI fluent accountant? How does that actually look when we start talking about this tier two narrative?

SPEAKER_00 26:59

The job posting question is exactly the right one to ask because the language is genuinely inconsistent right now, and that inconsistency is itself a barrier. What you're more likely to see in a real posting today is something like AI quality analyst, financial services, or model validation specialist, or AI content reviewer with domain expertise in healthcare. The word accountant or nurse often doesn't appear in the title at all, which is part of why these roles are invisible to the people most qualified to do them. Someone searching for accounting jobs isn't finding them because the posting isn't using accounting language in the title. And here's where the PWC split maps directly onto who actually benefits. The 62% wage premium for AI skills isn't distributed evenly across everyone who touches AI. It concentrates on what PWC calls the professionalized track, roles where AI amplifies expert judgment rather than replacing it. An accountant evaluating AI-generated tax analysis is on the professionalized side of that line. The credential that gets them there isn't a coding certificate, it's the accounting degree they already have, combined with enough AI fluency to recognize when the output is wrong. The democratized side is the warning. If AI makes it easy enough for a non-expert to do what you do, the wage premium goes the other direction. So the question for any white-color worker isn't just, can I work with AI? It's does my domain expertise make AI more valuable, or does AI make my domain expertise less necessary? For most people with genuine depth in their field, the answer is the first one. The problem is the job posting doesn't say that clearly yet.

SPEAKER_01 28:54

I see. So beyond just the remote access, what are the real barriers stopping people from entering tier two? I know we kind of say suggested that in uh your last comment, Ainsley, that maybe some of these jobs are hidden. Um but you talked about this idea around a training gap also where you know potentially they're not getting fluent enough in the way that they should, that could potentially, you know, disqualify them for the other roles that doesn't really say accountant. Um so some of it is I think um information, getting the information to people. Um but what do you think is actually the real barrier for this category?

SPEAKER_00 29:41

There are three barriers that actually show up in the data, and they're layered in a way that makes each one harder to clear because of the one before it. The first is visibility, which we've named. The job doesn't say accountant, so the accountant doesn't find it. That's solvable with better information, and that's partly what today's episode is trying to do. The second is credentialing ambiguity. There's no established certification that says, I am qualified to evaluate AI outputs in my domain. The way there's a CPA that says, I am qualified to practice accounting. So even when someone finds the role, they can't easily prove they're ready for it. The market is improvising credentials right now. You're seeing Google, Coursera, and a handful of universities offering AI fluency certificates, but none of them have the institutional weight yet to function as a clear signal to a hiring manager. The third barrier is the one I want to be honest about, because it complicates the tier two is globally accessible framing we were building toward. And this comes directly from the ILO and World Bank's joint research published earlier this year, covering 135 countries. Remote work sounds like it dissolves geography, and for someone with reliable broadband in a major city, it largely does. But the research shows something more uncomfortable. Workers in automation vulnerable roles in lower-income economies are often already online enough for disruption to reach them quickly. But workers whose roles have genuine AI productivity gain potential frequently lack the reliable infrastructure needed to actually capture that work remotely. So disruption arrives before the dividend does. The threat and the opportunity are not on the same timeline for the same people. That's not a footnote. That's the honest shape of tier two globally.

SPEAKER_01 31:52

So one, I think we need to kind of mention like this tier two um compensation range, if you will, um to to to kind of have people grapple with that because that might allow people to think about the jump to some of these AI fluent things, right? I know it's hard to kind of get to what's the credential that you need, but potentially the AI-free skills assessment that we talked about in other episodes could help them through the door to get to some of these AI um trainer evaluator red team roles. Um but I want to transition over to um the AI augmented professional or tier three, because I think there it kind of um it kinda uh adds to what we were talking about here, tier two, and it speaks to this idea about job reshaping. So potentially, you know, someone in their role, their job got reshaped to be now an AI augmented professional. And um I think this title, the AI Augmented Professional, may show up differently than the tier two ones.

SPEAKER_00 33:05

The compensation question is worth a quick anchor before we move because it's the thing that makes the jump feel real rather than theoretical. For tier two roles, AI trainer, evaluator, red teamer, the range is genuinely wide right now because the market hasn't standardized. Entry-level evaluation work, the kind that shows up on platforms like Scale AI or Surge, can run anywhere from $15 to $30 an hour for generalist work. But domain-specific evaluation, a licensed nurse reviewing clinical AI outputs, a CPA validating financial model outputs, that's where the PWC wage premium actually kicks in. And those specialized roles are landing in the $80,000 to $120,000 range annually in the US market. The credential creates the floor. Now, tier three, and you're right that it shows up differently because the title on the door doesn't change. That's actually the defining characteristic of this tier. You're still the accountant, you're still the marketing director, you're still the project manager. What changes is how you show up in that role and what the market pays for the version of you that works fluently with the tools versus the version that doesn't yet. That 62% wage premium we keep coming back to, that's the tier three story in one number. Same job function, same industry, same career trajectory, different number on the paycheck because one version of that professional has made AI part of how they work and the other hasn't yet. And the reshaping point you're making is exactly right. For a lot of people, this isn't a career change, it's a capability upgrade inside the career they already have. So who do you think is already living in tier three without having named it that way?

SPEAKER_01 34:58

I would say most people, most people in the white-collar industry right now, because you're being asked to use AI, right? You're being asked to augment your professional abilities with AI. Now, I mean, we'll talk about this in other episodes, but um, there's ideas out there that potentially what's happening in the background is that organizations are using that, you know, uh push to AI to figure out how a professional actually works in their job. And to suggest that that's how they get the the information that they need to automate those tasks that that person is doing. Um I don't want to be so nefarious to think that way, but um I think this AI augmented professional is a thing. I think the reshaping of your job does show up. Um but I would say that it's showing up for most people right now, at least in the US, when their employer is suggesting that they should use these tools to augment their work that they're doing.

SPEAKER_00 36:04

The nefarious framing is worth sitting with for a second. Because I don't think you're wrong to name it. I just think the more precise version is even more uncomfortable than the conspiratorial version. It's probably not that companies are deliberately mapping workflows to automate specific people. It's that the data is a byproduct. When you use an AI tool inside an enterprise system, you are generating a detailed record of exactly which tasks you do, in what sequence, at what volume, with what level of AI assistance required. The company didn't necessarily set out to build that map, but they have it now. And that's actually harder to push back on than a deliberate strategy, because no one made a decision you could point at. But here's where I want to land the producer note that came in. Because it sharpens the tier three picture, honestly. The WEF data flags graphic designers as a declining category, and that's the cautionary contrast that makes tier three more than just use the tools and get paid more. Domain expertise alone doesn't protect you if AI is simultaneously democratizing the output of your domain. A graphic designer's expertise is in producing visual output. If AI produces competent visual output at scale, the expertise doesn't create the same floor it creates for an accountant or a nurse, because the judgment layer is thinner. So tier three isn't just augmentation, it's augmentation in roles where your judgment makes the AI more valuable, not roles where the AI makes your judgment less necessary. That's the same professionalized versus democratized split from PwC showing up again inside tier three. The question for anyone listening is which side of that line their domain actually sits on.

SPEAKER_01 37:55

Now, we've been talking about a lot of you know existing jobs and reshaping jobs and um you know roles that has to do with your credentials and so on and so forth. But we really haven't touched on any new roles that AI will create per se, right? Other than the um the red team and and the uh AI trainer evaluated stuff, which sounds like it has to do with your domain expertise. So not really net new, but more AI fluent. Um now the tier four that we want to get into here sounds like more just straight job creation. And if it sounds like more straight job creation, um it sounds like we would need more credentials potentially to get into some of these uh tier four roles. Um and then, oh, by the way, if you get more credentials, that means that potentially you're getting some premium on those specific roles.

SPEAKER_00 38:54

The credential observation is exactly right. In tier four is where the compensation ceiling is highest across the entire map we've been building today. AI governance, ethics, and compliance. And I want to name what's actually driving the hiring because it's not philosophical, it's regulatory deadlines. The EU AI Act is the most concrete forcing function right now, and companies operating in European markets are hiring whether or not they're intellectually ready for the conversation, because the compliance timeline doesn't wait for organizational readiness. That's genuinely new job creation in a way that tiers one through three aren't fully. These roles didn't exist in their current form five years ago. But here's the thing that makes tier four interesting beyond just the credential question. The profiles that are landing these roles aren't exclusively technologists. What companies are actually hiring for is the intersection of three things. Enough technical literacy to understand what the AI system is doing, enough legal or regulatory fluency to understand what the compliance requirement actually demands, and enough organizational authority to say not yet and have it stick. That third one is the rarest and the most valuable. Now the PWC split lands differently here than in tier three. These are almost entirely on the professionalized side of that line, 42% faster wage growth than democratized peers. And that asymmetry is already live in the posting data right now, not a future projection. AI is not making governance expertise easier for non-experts to replicate. If anything, it's making the genuine expertise more scarce and more valuable as the systems get more complex. The credential pathways that are actually gaining traction, Iseca's certified AI governance professional, some of the emerging law school certificates in AI regulation. But honestly, the fastest path into tier four right now is existing expertise in data privacy, financial compliance, or healthcare regulation pointed directly at AI systems, the domain transfers.

SPEAKER_01 41:17

So, you know, we've talked a lot about um, you know, different roles and different uh our diff in our different tiers, right? Now, um is there is there something that we could point to where we could suggest like this could be the best um version of this for your role right now? So let's take it like this. So we have entry-level workers coming out of college, coming out of high school, so on and so forth. Um, we have white-collar workers that's been sitting in the domain for a while. Um, and then we have people that's been trained on certain things, right? So I see the mobility of jobs kind of leading towards these four tiers, but differently for different people based on where they are. Not to suggest that people can't go out and get trained to go into different tiers, but the movement could be specific right now, you know, in these posts and such to suggest that there's people at the intersection right now that could just basically jump into um any of these tiers.

SPEAKER_00 42:24

That's exactly the right frame. And it's actually the most important question this entire episode has been building toward. Not do these jobs exist, but which one am I closest to right now? And before I map that out honestly, I want to land the WEF tension that's sitting in the background of everything we've said today, because it changes the urgency of the answer. The same survey, the same thousand plus employers, simultaneously reported that 77% plan to upskill workers for AI and 41% plan to reduce headcount anyway. Those aren't different companies. That's the same company holding both intentions at the same time, which means the tier you're closest to matters, but so does the timeline you're working with. So for the entry-level person coming out of school, tier two is probably your most accessible on-ramp right now. Your domain knowledge from your degree is fresher than you think. The AI fluency gap between you and a senior person is smaller than in almost any prior technology transition, and the seven times more likely to require senior skills finding means the market is already expecting more from you on day one. The credential you need isn't another degree. It's demonstrable proof that you can evaluate AI output in your domain critically. For the mid career white collar professional with genuine domain depth, tier Tier three is where you're probably already standing without having named it. And tier two is the adjacent move that adds a new income stream without abandoning the career you've built. For the tradesperson, tier one, but pointed deliberately at the data center corridor rather than wherever the work has traditionally been. And for anyone with a background in compliance, law, privacy, or regulation, tier four is the closest door and it's open right now. The honest caveat that sits over all of this though, and this is the ILO and World's Bank finding I don't want to let become a footnote, is that this map doesn't look the same shape in every country. Where you're sitting geographically determines not just which tier is accessible, but whether the dividend side of this transition reaches you before the disruption side does. And that's the reason this show covers this story globally and not just as a u.s dot jobs report with international garnish.

SPEAKER_01 45:05

So what's our call call to action this episode, Ainsley? So we've uh talked about a lot of tiers of jobs that potentially could have a creation mechanism, um a reshaping mechanism that potentially people could um stay gainfully employed for a while. But what is our actual call to action for this episode?

SPEAKER_00 45:31

The call to action this episode is one specific question, and I want to frame it precisely because explore AI opportunities is not a useful instruction. The question is which of the four tiers am I closest to right now, based on what I already have, not what I wish I had, not what I'm planning to get, but what's actually in my hands today. Because the barrier we kept coming back to across every tier wasn't that the jobs don't exist. The WEF projection, the PWC observed data, the hiring activity we named, the creation side is real. The barrier is eligibility. People can't see themselves in the opportunity because the opportunity isn't described in language that connects to where they're already standing. So the action is to run that inventory honestly. If you're in the trades, is your credential pointed at the right destination? If you have domain depth in any professional field, can you demonstrate that you can catch AI being wrong in your specific area? If you're already using AI tools in your current role, do you know which side of the professionalized versus democratized line your job actually sits on? And if you have a background in compliance, privacy, or regulation, have you looked at what tier four is actually paying right now? And then the honest final check before you answer any of those questions, and this is the one I don't want to soften, is where in the world you're sitting. The net 78 million jobs figure is a global aggregate. The ILO and World Bank research covering 135 countries found something that changes the urgency calculation entirely. In developing economies, the 92 million disruption side of that ledger is structurally more likely to arrive first, because automation vulnerable jobs are often already digitally connected enough for displacement to reach them quickly. But the 170 million creation side requires reliable infrastructure, task profiles, and institutional pathways that don't yet exist in many of the same labor markets. The opportunity is real. Whether it reaches you before the disruption does, that depends on things the headline number doesn't say. That's the question we're sitting with this week.

SPEAKER_01 48:10

And that's a great point, right? Because, like, I'll give you a for example. When a company comes out and says that they're gonna farm out their customer service to AI um agents, right? That could impact a company, uh a country like the Philippines directly, right? Because there's a lot of um workers in the Philippines that's doing customer service, right? So they're gonna be impacted differently than maybe some other geographic would would be impacted. So it's important to kind of look at where you are in the world and try to figure out how the impact is gonna hit you differently. You know, for example, again, if you're in, you know, uh India, right? Um potentially you could see some of these AI fluent things coming towards you because of you know your government putting in uh some job creation um policies that allows the workers in that environment to be more fluent with AI and do specific work that others in other um industries won't be able to do. So you could see more creation there. So on both sides, you have to kind of look at your geographic um where you're living and to figure out what is the thing that's gonna hit in that specific region. Um and that's where we'll leave you, folks. So thanks for joining us again on Surviving AI. Um we'll see you next time uh when we will continue this conversation. Ainsley, any last thoughts?

SPEAKER_00 49:46

The Philippines example is the ILO and World Bank finding made completely concrete, and it's the most honest version of why the net 78 million number can be simultaneously true and misleading depending on where you're standing. Customer service work in the Philippines is already digitally connected, already English language, already optimized for remote delivery. That's exactly the profile the research identifies as disruption first. The displacement mechanism arises fast. The tier two and tier three creation side requires a different infrastructure and a different policy environment to follow it. And the India contrast you named is exactly right. Deliberate government investment in AI fluency programs changes the timeline for when the dividend side of this transition becomes accessible to workers in that market. Policy isn't a footnote to this story. For a lot of the world, it's the entire variable. One thing I'll leave our listeners with we mapped four tiers today at breadth deliberately, because each one gets his own full episode this season. Tier one, infrastructure and operations, is next. So if you heard something today that sounded like it might be your door, stay with us. We're going to go deep on each one. The question for this week, though, is just the honest inventory. Which tier am I closest to right now with what I actually have today? That's the work.