The Governance Gold Rush: Who Gets Paid to Say No to AI

Somebody in your building, probably in privacy, legal, or IT, is about to get handed a job that didn't exist two years ago. This week, Carlo and Ainsley open Season 6's final Opportunity Map tier, AI governance, with the number that carries the whole episode: in a survey of 671 organizations across 45 countries, 98.5% say they need more AI governance staff in the next 12 months. Hiring has held flat at roughly 71 new US postings a week since January, including straight through the week everyone calls "the EU deadline," with no spike.

That flat line is the thread Ainsley pulls hardest. What actually activated August 2, 2026, was the EU AI Office's enforcement power over general-purpose AI models and Article 50's transparency rules, not the broader high-risk system regime, which the Digital Omnibus pushed to December 2, 2027. This is the third time the show has corrected that exact date on air. Carlo and Ainsley trade real (composite, disclosed) stories along the way: a privacy manager in Ontario whose title changed without a single job posting, a compliance team in the Netherlands that sprinted for a deadline that wasn't the real one, and a risk analyst in São Paulo who used a consulting job as the door the in-house market wouldn't open for her. The honest breakdown of who's actually getting hired: 22% privacy, 22% legal and compliance, 17% IT, and over 60% of governance leadership coming from three functions that have never touched a line of model code. Professional Services firms, not tech companies, are the single biggest hirer, at 35% of all postings, and the concrete, less-hyped ISO 42001 audit track pays $95,000–$140,000 in-house.

This closes the four-tier Opportunity Map that opened Season 6: infrastructure and operations, AI training and red-teaming, AI-augmented professional work, and now governance. Carlo and Ainsley are honest that the math doesn't fully add up: four real, hiring-now doors are nowhere near the World Economic Forum's much larger jobs-created projection, and the entry-level rung is getting thinner for people with no adjacent experience to redirect. But for anyone already doing risk, compliance, or process work today, the door is open now, wider than it's likely to stay. Wednesday: the tactical, 90-day version of walking through it.

00:00 Intro — Three Functions Now Run AI Governance
01:44 98.5% of Companies Say They're Understaffed
03:58 The Privacy Manager Who Became the Accidental Governance Chief
05:51 71 Postings a Week, No EU Deadline Spike
09:26 The Third Correction: What Actually Activated August 2nd
11:06 Who Has the Standing to Say Stop
13:21 The ISO 42001 Audit Backlog
16:55 The Netherlands Company That Prepped for the Wrong Deadline
19:07 The Real Cost of Panic Hiring
20:37 Why Consulting Is the Fastest Door In
23:36 The 22-22-17 Breakdown
28:02 The Honest Gap — Four Doors, Not 170 Million Jobs
31:21 Get In First, Certify Second
37:59 No Computer Science Degree Required
40:11 Four Doors, One Question: Who Gets to Say Stop

Subscribe: Apple Podcasts — https://podcasts.apple.com/us/podcast/surviving-ai-job-automation-workforce-future-insights/id1864360631
YouTube — https://www.youtube.com/@SurvivingAIRisk/videos

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

Transcript

SPEAKER_02 0:00

Somebody has to sign their name on 70% not being good enough. And honestly, three separate signals all pointing the same direction? The IAPP staffing gap, the flat hiring pace, even Forrester's Fortune 100 projection, that's not hype stacking. That's a real pattern converging from three different angles.

SPEAKER_01 0:25

Artificial system online.

SPEAKER_00 0:49

So, welcome back to Surviving AI. This is our AI governance episode. And we're talking about the jobs. We're saying no to AI is the whole point. So there's three top functions that's involved in AI governance right now. And the leaders are privacy, legal, and IT. None of those three are genuinely experts in AI specifically as it pertains to model development. But they understand things around AI that is uniquely in tune to the governance part of AI, which are the risk, the process, and how to say no in a way that it sticks versus any other roles in the organization. And that should bother engineers at least a little because you know engineers always feel like they make the decision on what goes to production or not based on their um title or their role in the organization. But relative to AI, these roles are prevalent to this idea around AI governance.

SPEAKER_02 2:22

That at-risk claim isn't a hunch. There's a number behind it that should stop people cold. IAPP's survey published with Credo AI polled 671 organizations across 45 countries, and 98.5% of them said they need more AI governance staff in the next 12 months. Not might, need. That's basically every organization in the sample admitting they're behind right now, not catching up someday. And the engineer discomfort you're describing is real, and I think it's worth sitting in for a second, because the instinct in every tech org has been whoever builds it gets to ship it. Governance flips that.

SPEAKER_00 3:34

I know I wouldn't put my name on there, right? So I think the risk of AI going wrong or getting it wrong is top of mind for folks, and they need help, right? That's why this whole episode exists is because AI governance is actually a thing. To be able to discern what exactly is happening relative to AI. I mean, harken back to pre pr previously before AI. You could go talk to a team, they could go present their work, and you could like investigate human-to-human what exactly did someone build, look at the technical details, and try to figure it out. But relative to AI and the black box nature of AI, no one knows what the heck is going on inside of it and where it may go wrong. So the governance part of it is the guardrail to say that I don't know what the risk is, we can't prove it, we can't show it. This can't ship.

SPEAKER_02 4:37

That black box point is exactly why this isn't a paperwork job. You can't interview an algorithm the way you used to interview a team lead. There's no meeting where the model explains its reasoning to a room and you catch the flaw in real time. Picture a privacy manager at a mid-size Ontario company. Someone who spent years doing exactly what you're describing, asking teams to walk her through what they built, poking holes, signing off. Her company rolls out a customer-facing AI tool, and she realizes she can't do that anymore. Nobody can walk her through it line by line. So she becomes the first person in that building with the actual authority to say, we don't understand this well enough to ship it. And that role didn't exist eight months earlier.

SPEAKER_00 5:24

And I'll push back on you a little bit, uh Ainsley. It did exist, right? Because the people that we mentioned before, that's in uh privacy, legal, and in IT, some of those folks was actually pushing pushing back on projects that had nothing to do with AI. So those folks kind of understand the inherent risk of um what a tool could do to our organization, or at least being able to capture what and scrutinize what the output of those tools are and come up with some sort of risk profile on what that tool is to suggest that okay, this is okay, let's ship it, or this is not okay, go back to the drawing board and put some other things inside of it. Right? So it's not as easy as to say that the role didn't exist. What we're trying to mention is that people have experience with some of the nuances of risk in general, and those people are front and center relative to AI governance.

SPEAKER_02 6:33

The skill set isn't new. The title is what she was doing at a generic vendor risk review is the exact same muscle, just pointed at something that doesn't explain itself the way a person does. So the honest version is the job didn't require new expertise, it required the same experts to get pulled into a new room. And that actually lines up with something in the hiring data that I think people keep getting wrong. Axial Search has tracked about 71 new US governance postings a week, every week since January, flat. No spike around August 2nd, no seasonal dip in the summer. If this were really about companies scrambling to meet a deadline, you'd expect a spike right around the EU AI Act date everyone keeps citing. There isn't one, which tells me this isn't the fire drill, it's these exact privacy, legal, and IT people getting permanently reassigned into a function companies have decided they need, whether or not a regulator is watching.

SPEAKER_00 7:35

And that's the right read, I would say, right? Because I mean a regulation shouldn't be the thing that um corporations are holding on to to be able to make sure that the thing that they're using, right, is having the right outcome so that people could take advantage of it. Right. The problem is that, you know, and you saw it from the laps last episode with the AI four, is that people don't understand it well enough to be able to discern whether or not it's right for them or not. Okay? There's so much nuances to it, there's so much chaos to it related to AI in general, that you know, companies are trying things out and they're not shipping because they have no idea what the heck is going on. So some of these frameworks that we have inside of the opportunity map season is what would help a company utilize this idea around a human in the loop to make sure that whatever AI projects that they're coming forward or pushing forward actually ends up at the right place. Right? Because without the human in the loop, at least in the short term, okay, a lot of this stuff falls flat. Like I said, without AI governance in an organization, okay, you're putting your organization at risk of bad outputs, bad outcomes relative to using AI. So human in the loop kind of helps the adoption curve such that you develop the muscle that you need in an organization to be able to put these projects forward. But it involves a lot of rigor, right? It involves a lot of new roles that you have never seen before. And that's what this season is all about. It's all about these opportunities that AI is creating relative to humans that would show up to be able to discern whether or not we're putting ourselves at risk. Are we building the right thing? Do we have the right risk profile? Do we have the right criteria? Are we answering the right customer questions? Is there an inherent bias here? A lot of different things they're infusing inside of um these new roles that we're talking about, right? Such that people could discern whether or not the tool that we're using is actually giving us the right output.

SPEAKER_02 10:05

There's actually a specific correction buried in what you just said, and it's worth being precise about it because this show has had to fix this exact date three times now. Once back in July, once on the AI-4 recap. What actually activated on August 2nd wasn't the AI Act deadline as one big undifferentiated thing. It was the EU AI office getting real enforcement power over general purpose model providers, plus Article 50, the transparency rules, chatbot disclosure, deep fake labeling, that kind of thing. The much bigger piece, the high-risk system rules that would cover hiring tools, credit scoring, medical devices, got pushed to December 2nd, 2027 under the digital omnibus. So companies that think they've already cleared the compliance bar because August 2nd came and went haven't actually faced the hard part yet. And that's exactly why your human in the loop point lands harder than a regulatory panic story would. If this were just about beating a deadline, you'd expect governance hiring to relax now that August 2nd is behind us. It won't, because the actual driver is what you're describing. Companies not trusting their own black box enough to ship without a person in the chain who can say stop. Which raises the question I think we should sit with going into who's actually filling these roles. If it's not regulation forcing this and it's not engineers being asked to do it, who has the standing inside a company to be that stop button?

SPEAKER_00 11:44

And that's a right question, Ainsley, that I expect you to say. Outside of just a demo that somebody showcase a thing in a narrow view, okay, have you really tested this thing a thousand times and every single time is right? Or is it ninety-nine percent? Is it eighty percent? What percentage has you have you tested this thing to to be able to sign your name and say I'm responsible for this thing, this thing now? Right? So I say the middle manager's has to take ownership here, and they have to be able to push on leadership enough to suggest that we're gonna go forward with a thing, or we're gonna just stop right now and try something else.

SPEAKER_02 14:00

That's the accountability gap nobody wants to name out loud. The person funding it isn't equipped to evaluate it, and the person closest to the technical reality often doesn't have the standing to say stop, which is exactly why the unglamorous version of this job is turning out to be the most concrete one. Not the flashy AI ethics officer title, the ISO 42001 audit track. It's an actual international management system standard, published back in 2023, and right now certification bodies like BSI and TUV are running six-month-plus backlogs just to get an audit scheduled, because there aren't enough qualified auditors to meet demand. In-house lead auditor roles are paying $95,000 to $140,000 in North America. That's not a hype number. That's someone getting paid to build exactly the have we tested this a thousand times rigor you just described on paper in a way that survives a regulator asking to see it. And here's the thing that should reassure your middle manager, not intimidate them. That auditor role isn't going to engineers, it's going to the person who already knows how to build a documented process and make it stick, which loops right back to where we started.

SPEAKER_00 15:22

So I want to talk about one point for a second, right? So we we've we've discussed the EU AI at nauseum on the show, right? And I I think it's the right frame, just generally. Softened, not softened, it doesn't matter. Yeah. The idea is that all of these roles that we're talking about in season six relative to AI and the right outputs, is necessary regardless of the EU AI Act. And that's why you kind of see this idea of you know, the 71 um roles per day or per month, I forget which which one it is, uh coming out. Because all of these roles relative to figuring out exactly the right output, the right guardrails, the right red teaming, the right uh evaluating, training, um, is necessary for you to get this right as an organization. And for the audience just generally, those roles have somewhat abstain power to suggest that organization will always need these roles, at least in the short term, okay, until AI gets so good that it doesn't really need any guardrails or it has his own. I have no idea. Yeah, nobody does. But for right now, we need to be able to see someone, human in the loop, looking at this black box and saying, okay, we had you know hundred thousand uh customer service calls yesterday, and the ratings on every single one of them is you know eighty-nine, ninety percent, whatever it is, okay. And what are those ten percent of customer service call that went off the rail? Can we go look at them, see what happened, and figure out what's wrong with the tool? Right? And even before you put the customer service tool in place, right, do you have the right framework in your organization backed by a human in the loop to suggest that someone signed off on this and say I risk my career on saying that this thing is ready to work.

SPEAKER_01 17:33

Quick fix on the number, since it's the one people misquote most, it's 71 new postings a week. And it's been flat at that pace since January. No dip, no summer lol. That flatness is actually the whole argument.

SPEAKER_02 17:49

If you want proof this is structural and not regulatory panic, that's it. And I'd add a concrete version of the cost of getting that date wrong, because it's not just a trivia point. Picture a logistics company in the Netherlands that built its entire AI rollout calendar around August 2nd as the deadline, as if everything hit that day. Teams pulled all-nighters, leadership froze half their AI projects out of caution, and then the actual high-risk rules they were bracing for don't apply until December 2nd, 2027. They panicked for a deadline that wasn't the one in front of them yet. That's the real cost of getting the date wrong. Not a footnote, an entire company running scared of the wrong finish line, which actually strengthens your point rather than undercuts it, because the conversion of that same company would have kept building governance capacity at a steady pace the whole time. The same way the 71 a week number has held steady instead of lurching round a date that turned out to be only half the story.

SPEAKER_00 18:55

I still think those companies are right anyway, Ainsley, right? Because being more prepared to a moment is better than not being prepared at all, in my opinion. Right? So even though the dates change, regulations change for whatever reason they did, right? I still think those companies are better off than that. So race to the finish line.

SPEAKER_02 19:56

You bring in people to hit a deadline that turned out to be the wrong deadline, and now you've got budget and headcount locked into a shape that doesn't match the actual 12-month tier or the actual high-risk rules landing in 2027. That's not free. That's a company that now has to re-architect a team it just built under a leader who already spent the political capital to get it funded once. But you're right that having the people, even imperfectly matched, beats having no one. And that's actually where the real entry door is, not the one people assume. Think about a risk analyst in Sampolo who applies directly for governance roles and keeps getting passed over, no track record. So she goes a different way, gets hired by a consultancy instead, and her actual job becomes building AI governance programs for other companies' clients. Professional services firms are 35% of all governance hiring right now, more than double tech or financial services. Two years later, she's not applying with the resume that says, I want to do this. She's applying with three built programs behind her name. And now the in-house job finds her.

SPEAKER_00 21:18

So this idea around consultancy relative to governance and some of the roles that we're talking about is actually a good one to pull on because the consultancy folks are actually seeing a lot of projects, right? So the thing that we mentioned in AI 4 the AI 4 episode was that, you know, if you're in an organization and you're doing some of these things, potentially what could happen is, you know, an organization doesn't see the value of whatever you're building and they shelve the project. And when they shelve the project, the people get caught up in the middle of that. Um in a consultancy, you kind of move from project to project to project, so it doesn't really matter, right? You already got paid. From a the consultancy side to say, let me go consult with some organization on whatever it is that they're looking for. So it is a good thread to pull on, right? Um for folks to look at consultancies with their domain experien uh experience to be able to get into some of these things that we're mentioning, right? And especially governance, for example, right? So if you're in policy, if you're in um legal or IT, potentially one of the big consultancy firm could hire you and say, hey, you know the rules, you know this come customer, you know that specific thing. Hey, come on over and um help us with government AI and making sure that the output of this tool is actually correct.

SPEAKER_02 22:47

That project portability point is actually the underrated advantage. You're not betting your whole career on one company's AI initiative surviving budget season. If the client kills the project, the consultancy just staffs you onto the next one. The risk that sinks an in-house hire, the pilot getting canceled, and no one remembering to find you a new seat basically doesn't exist in that model. And it maps exactly to that 35% number. Consulting and advisory firms are the single biggest employer of AI governance talent right now, more than tech companies, more than financial services combined. Which tells you something people miss. The earliest, most flexible door into this whole field isn't a job posting that says AI governance lead. It's a firm that already has the client relationships and just needs someone who actually knows privacy or knows healthcare compliance or knows financial regs to go be the translator on site. Which is actually a clean way to zoom out on where we've landed today. We opened with a staffing gap that's basically universal, 98.5% of organizations admitting they're shorthanded. We corrected the record on what the EU actually activated versus what's still two years out. And now we've mapped the doors in privacy, ISO auditing, consulting, domain expertise, none of which run through a computer science degree.

SPEAKER_00 24:15

So if it doesn't run to a computer science degree, who actually gets hired for some of these roles, right? I know we mentioned uh privacy legal complies IT, maybe, right? And additionally, what's the on-ramp for some of these folks to get into it? And oh, by the way, don't forget about entry-level workers and the idea that so you know some people are retiring, we got entry-level workers waiting, and they almost have no door to show up at.

SPEAKER_02 24:47

The honest breakdown is 22% privacy, 22% legal and compliance, 17% IT, that's the IAAPP survey again, and it's not close. Combined, that's over 60% of governance leadership coming from three functions that have never touched a line of model code. The on-ramp for those folks is fast by tech standards, six to twelve months, because they're not learning a new discipline, they're redirecting a discipline they already have. Someone who spent eight years doing vendor risk review or privacy impact assessments isn't starting from zero. She's starting from, I already know how to write a risk memo that survives a lawyer's read and just needs the AI-specific vocabulary layered on top. Entry level is the harder, more honest answer, and I don't want to soften it. The guides put fresh graduates at 12 to 18 months, not six, because you can teach someone the framework, but you can't shortcut the domain judgment that comes from having actually sat in rooms where something went wrong. And there's a real tension underneath that. The same efficiency argument that's making companies want fewer, more senior hires everywhere else is going to make that door narrower right when young workers need it most. So the honest move for someone starting from nothing isn't a governance title on day one. It's an analyst seat at a consultancy that's staffing people onto governance engagements, riding alongside the senior person, and building the portfolio the direct hire door is asking for. Which actually closes something bigger than just this episode. Infrastructure, training and evaluation, AI augmented work, and now governance. That's the whole map this season set out to draw, door by door. And what strikes me, sitting here at the end of it, is that none of the four doors required starting over. Every single one was take what you already know and point it somewhere new.

SPEAKER_00 27:03

So let's talk about that a little bit, right? Because um we talk about entry level and we talk about some specific roles in um white collar work and with the compliance, the policy, the IT, and legal. But what about every up everybody else, right? The the thing that I'm stuck with is that this idea that you need to almost find a ha a home for everyone that could get impacted by AI, right? Uh, we started the season off with this idea that you know the WEF World Economic Forum was basically mentioning that 95 million jobs would be re would would be impacted and 170 million would be created. I'm trying to find the 170 million in what we're talking about this season, the opportunity map. I think some of them is in there. Okay. When you start start talking about um people pivoting into trades, people going into other fields like data center operations or fiber technician and some of these other things that we talked about. And then similarly with the AI red team and um evaluating and um training, and then now with the governance and AI augmented individual. Now, is that enough? That's really what I'm left with. Is that enough to kind of make sure that everybody have a place to go, right? If some of these um unemployment or some of these impacts from AI does in fact show up at the white-collar worker desk.

SPEAKER_02 28:40

That's the question I don't think this season has fully answered. And I want to be honest about that rather than paper over it. We've mapped four real doors now: infrastructure and data center work, training and red teaming, AI augmented professional roles, and now governance. Each one is genuinely hiring, genuinely paying well, and genuinely accessible without a computer science degree. That's not nothing. But if you add up the actual headcount across all four, even generously, you're nowhere near 170 million jobs. You're talking about hundreds of thousands of roles globally, maybe low millions at the outer edge, against a World Economic Forum projection measured in the hundreds of millions. And the honest gap is who those four doors don't reach. They all assume you're already an adult with an existing skill set adjacent to the work, a privacy professional, an electrician, a domain expert with 10 years in healthcare. Nobody on this map has answered what happens to the person whose actual job was the thing AI replaced outright with no adjacent skill to redirect. The International Labor Organization's own research flags this directly. So the truer version of the season's promise isn't here's where 170 million people go. It's here are four doors that are open right now for people close enough to walk through them.

SPEAKER_00 30:37

And oh, by the way, you have a certain amount of time to walk through through them, right? Because I mean these jobs are hiring right now with seemingly um domain experience, slash, you know, you have some experience in this type of thing, come on over, right? But ultimately that might change in the future, right? They might be more rigorous to say that, hey, now I want a couple of years of experience with doing this thing or not. Right? So if these jobs are really the ones that we'll rely on in the future, you know, it behooves us to start at least trying to get into some of these things right now. Now, I guess that's what I'm left with here, Ainsley, is like what is the actual credential to get into some of these things, right? I know we talked about in other episodes the AI or the GPAI um you know, credential um that you could go get and um try to use that to you know see if you could get into it. But I know a lot of these roles right now is not even talking about credentials in general, but how does folks get into these things that we're talking about now, you know, that that are hiring to be able to get into AI governance?

SPEAKER_02 31:59

Window Point is the urgent part, honestly. Right now, only about 12% of governance postings actually require a certification. That's not a typo. That's the whole opportunity. Companies are hiring for judgment and adjacent experience because a function is too new for anyone to have five years of AI governance on a resume. Nobody does. The job barely existed 18 months ago. But that's exactly the kind of gap that closes itself once the field matures. Give it two, three years, and posting starts saying prior AI governance experience required. And suddenly the people who got in during the adjacent experience welcome window are the only ones who qualify for that requirement. The door doesn't stay this wide by design. It stays wide because nobody's built the gatekeeping yet. On the actual credential, IAPP, the group behind that 98.5% survey we keep coming back to, runs the AIGP, the Artificial Intelligence Governance Professional Certification, and it's built specifically for people coming from privacy and compliance rather than engineering. IAPP's own data says certified professionals earn 13% more with one credential, over 27% more with two. And on the more technical operations side, there's the ISO 42001 lead auditor track, we talked about earlier. That one's less about title inflation and more about an actual auditable skill, which is part of why it's paying $95,000 to $140,000 before anyone's even negotiating. But the sequencing matters more than the credential itself. Get in first, certify second. Most of these hiring managers aren't screening for the AIGP today. They're screening for do you understand risk and can you make a no-stick? Wait for the perfect credential and you'll be qualifying for a door that's already narrowed. That's the real close on this season's map. Infrastructure, training, augmentation, governance. Four doors. And the honest news is they're open now, not open indefinitely. Wednesday we get into the 90-day version of actually walking through one.

SPEAKER_00 34:16

So let me push on something slightly here, Ainsley. This idea that you get the credential afterwards, that doesn't really always mean that you are gonna get some sort of pay rise or something with the organization that you you're working with right now. Essentially, um, based on what Ainsley mentioned there and the data, is that if you got if you have credentials, potentially you can make more. But if you don't have credentials and you get in, and you do get credentials after you get in, potentially you might have to find somewhere else that's gonna give you the premium that you're talking about. That's a good problem in my head, yeah. Because you get in gainfully employed, you get credentials, right? Still gainfully employed, and then potentially you're more marketable and marketable now because you have this idea this uh new credential that you could go shop around to everybody else that's looking for uh AI governance uh person, right? So that's what I would say. That's the only thing I would push back so far. But I mean, this whole season is the opportunity map, folks. So there's a lot of opportunity out there relative to AI, right? Not converse uh of AI, but relative to AI. Companies are trying to get to it. Um they're somewhat uh you know, not shipping everything that they're suggesting. But the moment the shipping part starts working, right, there's gonna be more impact that's happening, and you don't want to wait until that thing happens to be able to think about, oh, let me pivot, let me move, let me people are already gonna be starting doing that, right? The people that's getting impacted through this year, even could be getting involved in some of that stuff, right? Um it's kind of unlikely because we we heard about like these outplacement firms kind of shoving people back into the thing that they normally did. But if somebody is smart enough to kind of come up with this idea that like I'm not gonna go and go be a software engineer, I'm gonna do this other thing, and the other thing, oh by the way, it's governance, right? And then they get into those roles, and some of those people could be taking the jobs that you could potentially get into if you're impacted in the future because they got these tools working the way they want it to.

SPEAKER_02 36:40

It's not really a rebuttal, it's the sharper version of exactly what I was gesturing at. The realistic play usually isn't get the AI G P and walk into your current manager's office for a raise. It's get in, do the work, get certified, and treat that credential as something you cash in with a different employer entirely. Internal comp cycles move slow and are political. The external market doesn't care about your history at that company, it just sees certified, six months of real governance work behind them, and that's worth more elsewhere than it is to the people who already have you. So the premium is real, it's just usually collected at the next door, not the one you're standing in. And that competitive dynamic you just described, other people racing into these same roles before displacement hits them, is actually the piece that makes the timing argument urgent rather than theoretical. Because the honest tension I keep churning over that I don't think this season fully resolves is the International Labor Organization's own research on this. AI isn't necessarily wiping out entire job categories. It's compounding how hard it already is for people early in their careers to get a foothold anywhere adjacent to these growing functions. So the same senior privacy person who pivots in six months is, without meaning to, closing the door a little further behind them for the 22-year-old who doesn't have eight years of vendor risk review to redirect. That's not a flaw in the map, it's the honest edge of it. The doors are real, they're just easier to walk through the closer you already are to them. And that's probably the right place to leave the season's four-tiered arc. Infrastructure, training and evaluation, AI augmented work, governance, genuinely open doors, genuinely time limited, and genuinely not evenly reachable for everyone standing in front of them. Wednesday is the 90-day version of actually stepping through one.

SPEAKER_00 38:37

And oh, by the way, no computer science degree, right? So a lot of people think about AI and they're thinking about like, I need to be a PhD, I need to be computer science, I need to know a model, I need to know Python, I need to know all of these things. No. What we're saying is that you need to know what you know, and then you need to be able to trust yourself, use critical thinking. We talked about critical thinking in episode in season five, use critical thinking and judgment to be able to discern that this output of the tool is not right, uh, and here's why. Okay, let's fix it before we ship it. Governance is all about that. Governance is all about looking at the risk, right? Is it uh infrastructure AI, is it uh employee AI, is it customer AI, and figuring out what's the impact to my organization if I ship this tool into the public, into my organization. Right? Do I have any downside risk to that tool being shipped as is, or do we need more time to figure out and ask the right questions and make sure that this is solving the right customer problem for everyone that's a customer, right? Not without bias, not without this, not without that, to be able to ship it. So, like Ainzy said, this is the arc, right? The governance, the um AI augmented folks, the infrastructure, the training and evaluating that has more to do with the model than that. But all of these jobs, net new to AI, even the trades one, to be honest with you, right? People that's making so much money off of being an electrician for a data center didn't have that before data center for AI infrastructure. Okay, so take advantage of all of these roles. It's not gonna solve uh world hunger, it's not gonna give us our 170 million uh WEF jobs, but it does give us something at least in the short term.

SPEAKER_01 40:49

The no computer science degree point is worth sitting with for a second, because it's genuinely the opposite of what most people assume walking into this: 22% privacy, 22% legal and compliance, 17% IT.

SPEAKER_02 41:06

None of those are coding disciplines. What they share is exactly what you just named. The willingness to look at a confident output and say, I don't care how fluent this sounds, show me why it's right. That's not a technical skill. That's judgment under pressure. And it turns out that's rarer and more valuable right now than knowing how to build the model in the first place. And it's worth remembering this isn't just an EU story either. India's data protection rollout phases in this November, which means there's a second real compliance trigger outside Europe that's going to need exactly the same translation skill, on a completely different timeline. The patterns showing up everywhere risk meets deployment, not just wherever Brussels happens to be regulating that quarter. What actually strikes me, closing out on this, is that the through line across all four tiers this season wasn't really about AI at all. It was about who still gets to make the call. Infrastructure asked who builds the pipes, training and evaluation, asked who checks the model's work, augmented professional asked who directs the tool instead of being replaced by it. And governance asks the sharpest version of that same question: who gets to say stop? Four different doors, same underlying bet. The skill that survives all of this is judgment, not code. That's a real season arc, and it adds up in a way I wasn't fully sure it would when we opened with infrastructure back in episode one.

SPEAKER_00 42:49

Gotta trust the process, Ainsley. That's all that's all I gotta say. But no, I'm being honest. Um listen. I hope people are paying attention, right? Because, you know, coming out of AI four, the one thing I realize is that there will be a different way of working in the future. Right? For everyone. Okay. A lot of poke folks are working with these tools in their jobs right now and prompting. Okay, you're going to Gemini, you're doing um, you know, Microsoft co-pilots, you're doing whatever you're doing, you're working with these tools and you're doing prompting. But there will be more things relative to your specific job title, marketing, legal, sales, there will be more tools directed straight at the thing that you're doing. And the folks that we're talking about, at least in this season six, are the folks that are gonna make sure that those tools, those things that touches the customer, are actually the are actually working correctly. Right? It's not necessarily the augmented individual per se, you know, relative to um, you know, the whole thing that AI could the whole breadth of AI, right? But the jobs that we're talking about that's hiring right now, AI red team or evaluator trainer, um, the trade jobs, the AI governance jobs, right, those are the ones that's actually have roles out there right now. The augmented person, that's yet to be seen, in my opinion, but you need to become AI fluent in whatever role you are, just generally. But the roles that's hiring right now are the ones that we're talking about in the season. And we'll talk more about this ninety-day AI governance playbook, and uh noodle on whether or not we use that or we use some other episode, but I think it's a good topic to continue to talk about this governance arc because um governance is a pretty big one. No one really knows exactly what's going on inside of the AI output and how it comes up with what it comes up to. And if somebody has to say no because they realize that the thing that this the AI tool is coming up with is 70% right, and that's not good enough, then somebody has to sign their name uh uh uh uh on that, right? So that's that AI governance.

SPEAKER_02 45:30

Somebody has to sign their name on 70% not being good enough. And honestly, three separate signals all pointing the same direction. The IAPP staffing gap, the flat hiring pace, even Forrester's Fortune 100 projection, that's not hype stacking. That's a real pattern converging from three different angles. That's the whole job in one sentence. Worth trusting. See you all Wednesday.