A Conversation about the AI Values Test

[00:00:00] Imagine it is Tuesday afternoon. You're sitting at your desk and you're staring down this massive, just convoluted data entry project. Or maybe a brutal coding task that normally eats up your entire Wednesday and like half of Thursday. Oh, we all know that feeling. The dreaded end of week grind. Right. But you've been experimenting a bit lately. You figure out this really clever way to chain a few generative AI prompts together. And suddenly, the 30-hour task

[00:00:29] is completely finished. Wow. It's flawless. It is just done. So the question is, do you immediately jump on Slack or run into your boss's office to show off this massive efficiency win? Well, if you look at the data, and if you are like the vast majority of the modern knowledge workforce, the answer is absolutely not. Exactly. Not a chance. You are going to keep your mouth firmly shut, you'll hide the fact that you use the tool, and you'll just trickle that work out over the rest of the week.

[00:00:57] Yeah. And that hesitation, that instinct to hide your own innovation is the absolute center of what we are unpacking today. We are officially moving out of the AI hype phase, you know, that phase that dominated the last three years. Right. The era where the sheer processing power was literally the only thing anyone talked about.

[00:01:15] Exactly. But that era is over. We are now deep into the AI integration phase. And in this phase, the friction isn't in the silicon or the data centers anymore. The friction is completely in human psychology. It's in workplace anxiety. Which is such a fascinating pivot. Because for the past few years, the only question in boardrooms was, you know, how fast can we deploy this? It was just this frantic race.

[00:01:39] Yeah, total arms race. But the people actually doing the work, the ones figuring out that 30-hour shortcut we just talked about, they're asking a completely different question. They're asking, if I show you this efficiency, are you going to use it to help me, or are you going to use it to eliminate my job? And what is truly wild is that millions of workers are willing to completely walk away from their jobs based on that answer. Which is crazy to think about. But welcome to our deep dive. Today, we are exploring this massive silent shift in the corporate world.

[00:02:08] And we're basing this on a brilliant October 2026 analysis by Dr. Jonathan H. Westover in the Human Capital Leadership Review. Yes, it's a phenomenal piece. It's called the AI Values Test. Right. And our mission today is to figure out how AI has secretly become a test of employer character. So, okay, let's unpack this. Because whether you're an employee sitting there wondering if your company is actually going to support you,

[00:02:31] or you're a manager trying to navigate this without your best talent running for the hills, you really have to understand this test. You do. Because it completely reframes our relationship with the tech. The whole premise is that AI is no longer just a productivity tool. It's not just a faster laptop. Right. It's something entirely different now. Generative AI has become a mirror. The way a company chooses to deploy it instantly reflects their true underlying priorities about their workforce.

[00:02:59] And employees are looking into that mirror, and they're making massive life decisions based on what they see, right? Absolutely. But before we look at how organizations are failing this test, we have to look at the massive, almost comical contradiction in how workers actually feel right now. The data from Dr. Westover's analysis reveals a workforce that is deeply, deeply conflicted. Yeah, the contradiction is just staggering. It is. I mean, we really have to look at the GMP World at Work report.

[00:03:27] This was just released this October in 2026. It's a massive survey of 3,100 executives and 4,000 employees. And that's across six major economies, right? Like the U.S., U.K., Australia, France, Germany, Singapore. Exactly. So we are looking at a highly representative sample of the global knowledge economy. And the headline statistic from that report, I mean, it stopped me in my tracks. 58% of employees said they would turn down a job offer if a company's AI values did not align with their own.

[00:03:56] Which is huge. Just think about that dynamic for a second. You go through three rounds of interviews. You do the whiteboard test. You get the offer letter. The salary is right. But you literally walk away from the paycheck because of the company's philosophy on artificial intelligence. Yeah. And on the surface, that might sound like some sort of neo-Luddite movement. You hear that nearly six in 10 people will reject a job over AI, and you immediately assume people are terrified of the robots, right? Right. That's the logical assumption.

[00:04:24] You picture this workforce that just wants to go back to typewriters and physical filing cabinets. Yep. But then you look at the very next data point in the exact same report, and it completely shatters that narrative. Oh, this part is wild. Of the employees who are likely to look for a new job in the next six months, and by the way, that's a massive 62% of the workforce right now. So restlessness is spiking.

[00:04:47] But of that group, 73% said they actually want their next employer to embrace AI as a core part of operations. It creates this fascinating paradox. Wait, so workers aren't terrified of the robots taking over. Not at all. They actually want to work with AI. They want the cutting-edge tools. But they will walk away if it's done wrong. How does a company even begin to navigate that? Well, to understand the nuance, we have to pull in some context from a Pew Research Center survey from October 2024.

[00:05:17] Back then, 52% of workers felt worried about the future impact of AI, and only 36% felt hopeful. Okay, so more worry than hope. Right. But when researchers dug into the qualitative responses, like the actual human conversations behind those numbers, it became clear that the hostility wasn't directed at the software itself. So they don't hate the AI? No. Most people are just exhausted by the friction of modern knowledge work. They recognize that generative AI can remove the drudgery.

[00:05:45] I mean, they want a tool that can summarize an hour-long meeting in 30 seconds, you know? Oh, who wouldn't want that? Or something that drafts boilerplate emails or builds a basic presentation framework. They are desperate for that relief. Exactly. So if they love the tool, what exactly is triggering the anxiety? What are they afraid of? That's the big question. They are deeply uncertain about who ultimately captures the value of that efficiency. Let's go back to your opening scenario.

[00:06:13] If you use a chained prompt to do a 30-hour task in four hours, what happens to the remaining 26 hours of your week? Do you get to go home at 3 p.m. on a Tuesday? Right. Do you get to do high-level creative work that might actually earn you a promotion? Or does your manager just look at the dashboard, see you have extra capacity, and triple your workload without paying you a dime more?

[00:06:35] Or, the darkest timeline, the manager realizes that since you can do the work so fast, they only need one of you instead of three of you, and they lay off your two teammates. Yes, that is exactly the fear. The anxiety isn't about the capability of the tool. It is a profound fear of the person wielding the tool. I love the analogy we were talking about earlier. It's like wanting a high-powered nail gun because swinging a hammer all day is just ruining your shoulder.

[00:07:03] Oh, that's a perfect way to look at it. You desperately want the nail gun. But you don't want the sight foreman to take that same nail gun and use it to dismantle the house you're currently living in. That analogy isolates the mechanism of the anxiety so well, and it explains this massive perception gap that's currently tearing through the corporate world. What do the numbers look like on that gap? Returning to the GP data, 49% of executives report being highly confident in their organization's AI outlook.

[00:07:32] They look at projected efficiency gains and reduced costs, and they feel incredibly optimistic. Nearly half the C-suite. Yep, but only 27% of employees share that confidence. So the executives are essentially popping champagne while barely a quarter of the actual workforce feels secure. If they're looking at the exact same technological rollout, how is the experience so drastically bifurcated? It comes from a profound communication vacuum.

[00:07:59] The analysis highlights data from Gallup from June 2025, and it exposes the structural flaw in most corporate rollouts. Okay, what did Gallup find? They found that 44% of U.S. employees acknowledged their organization had started integrating AI into daily workflows. But of that group, only 22% reported that leadership had communicated a clear plan for how that integration would actually happen or, you know, what the end goal was. Yeah.

[00:08:26] So half the workforce has these powerful tools just dropped on their desktop, but only a quarter have been told why. Exactly. And if you are listening to this, you've probably lived this exact scenario. Over the last couple of years, you've seen these tools just silently appear in your software suite, like a little sparkle icon suddenly shows up in your email client. Yes, the magic wand button. Right. Or a new sidebar appears offering to rewrite your text.

[00:08:50] The deployment is incredibly fast, but the policies, the actual town hall meeting where the CEO explains what this means for the accounting department, that communication is lagging months or even years behind the software update. And, well, nature abhors a vacuum. When leadership fails to communicate the why and the how, they aren't just delaying a meeting. They are inadvertently forcing their employees to play detective. Oh, definitely.

[00:09:15] If the executive team isn't explicitly outlining their motives, the workforce has no choice but to infer those motives based entirely on leadership's observable day-to-day actions. And this brings us to the deep psychology of the workplace. Because employees are incredibly astute observers of corporate behavior, right? I'd argue they are hyper-vigilant right now. Oh, absolutely. If my boss isn't giving me a roadmap, I'm going to start reading the tea leaves.

[00:09:42] I'm scrutinizing who's getting promoted, who's quietly being let go, and what strange new metrics are suddenly on my performance review. And in organizational psychology, we have a very clear way of categorizing what you're observing. We make a critical distinction between stated values and enacted values. Okay, break those down for us. Sure. So stated values are the superficial declarations. They're the bullet points on a glossy PDF memo.

[00:10:11] They're the core principles painted on the wall of the corporate lobby. Like, we believe in human-centric innovation. Exactly. Or our people are our greatest asset in the digital age. Which sounds lovely. It's very comforting PR. But I imagine enacted values are much harder to fake. They are impossible to fake. Enacted values are the reality of what happens when decisions actually cost the company money, time, or leverage. They are revealed by the mechanics of the operation. Like what specifically?

[00:10:41] Like which specific tasks is the company choosing to automate first? When a department saves 40 hours a week using a new enterprise LLM, who captures that time savings? Do they get breathing room or is their headcount frozen? Right. Are frontline employees consulted before a new platform is bought? Or is it mandated from the top down? And crucially, how is liability handled? Ooh, liability. That's a huge one.

[00:11:06] If an AI tool hallucinates and an employee unwittingly submits that bad data in a report, who gets punished? The vendor? The manager who mandated the tool? Or the frontline employee? I mean, the employee is going to weigh those enacted values a hundred times more heavily than the stated values. Talk is cheap, but getting a formal reprimand because the mandatory AI hallucinated a number? That's a very real signal of how the company views you.

[00:11:33] To truly understand how workers process these signals, Dr. Westover's analysis grounds this dynamic in three robust academic frameworks. Let's get into those. These are psychological theories studied for decades, but they're suddenly hyper-relevant because generative AI acts as such a powerful catalyst. The first is signaling theory, originally developed by the economist Michael Spence back in 1973. 1973.

[00:11:59] That is decades before anyone was talking about chat GPT or neural networks. Long before. Spence actually won a Nobel Prize for this work. The core mechanism of signaling theory is about navigating situations with high uncertainty and asymmetric information. Meaning one party holds significantly more information than the other. Exactly. In these scenarios, the party in the dark doesn't just give up. They actively scan for observable actions that are costly or difficult to fake.

[00:12:28] And they use those actions to infer the hidden motives of the informed party. Let's ground that in a non-academic analogy. It's like playing high-stakes poker with someone wearing mirrored sunglasses. You can't see their eyes, and they aren't telling you what cards they hold. That's the asymmetric information. Right. So you don't look at their face. You look at the mechanics of how they bet. Did their hand shake when they pushed the chips in? Are they betting aggressively early on or waiting?

[00:12:55] You use their actions to infer the hidden reality of their hand. That is a brilliant application of the theory. In the corporate AI context, the CEO holds the cards. They know the three-year strategic plan. The employee's in the dark. So what happens when the CEO makes a move?

[00:13:13] Well, when a CEO announces a massive, multi-million dollar AI investment, in the exact same quarterly earnings call where they announce a freeze on all mid-level hiring, the workers read that choice. Yeah, that's a definitive signal. It is. The enacted value signals whether the organization views its employees as collaborative partners or simply as costly line items that need to be engineered out of the budget.

[00:13:39] That brings tremendous clarity to why a communication matters so much. If you don't show them your cards, they will guess, and usually they'll guess the worst case scenario. What is the second framework from the analysis? The second is known as person-organization fit. This stems from a massive body of research showing that human beings are deeply driven by a need for congruence between their own personal moral compass and the values of the institution they work for. So alignment is key. Yes.

[00:14:08] When an individual perceives a high degree of alignment, it acts as a massive driver for retention, job satisfaction, and discretionary effort. But when they feel a clash, the friction becomes unbearable. What does that clash actually look like in an AI context, though? Well, imagine an employee who has built their entire professional identity around craftsmanship, extreme accuracy, and client care.

[00:14:31] Now, imagine their company deploys an AI customer service tool with the explicit goal of handling a massive volume of tickets, even though the tool frequently gives slightly inaccurate or tone-deaf answers. Oh, man. So the company accepts the lower quality just for the cost savings. Right. And the employee is now forced to participate in a system that violates their personal value of quality. That perceived lack of fit creates an intense psychological dissonance, which inevitably drives turnover.

[00:15:01] Which brings us full circle back to that 58% statistic from earlier. If a candidate values transparency and quality and they sense during the interview that the company values opaque efficiency at all costs, they won't even take the job. Exactly. The fit is broken before day one. But it's the third framework in the analysis that I think is the most visceral, and honestly, the most emotionally charged for anyone who's been in the workforce for a while, the psychological contract. Yes.

[00:15:27] The psychological contract is a concept heavily developed by researchers like Robinson and Rousseau in the 1980s and 90s. And we have to be clear here, this is not the physical piece of paper you signed with HR. It's not your job description. No. The psychological contract is the unwritten, highly subjective, and reciprocal set of beliefs and expectations that exist between you and your employer. It's the implicit vibe of the workplace.

[00:15:52] You know, if you think about it in terms of personal relationships, it's like a marriage or a really deep friendship. How so? Imagine you and your partner have an unspoken agreement that you always split the household chores 50-50. It's just how you've operated for years. Then suddenly, they just stop doing the dishes. They completely stop doing laundry. Oh, that would be infuriating. Right. They expect you to handle all of it, but they never actually sat down and had a conversation with you about changing the dynamic. Technically, there's no legal document dictating chore division.

[00:16:22] They haven't broken a law. But emotionally, interpersonally, it is a massive structural breach of trust. The emotional weight of that analogy is exactly what employees are feeling right now. Now, Generative AI is currently rewriting the implicit terms of millions of psychological contracts every single day without any consultation. People are just left wondering what skills are actually valued here now. Exactly.

[00:16:47] They're asking, does my historical knowledge of this company mean anything if the AI can retrieve the data faster? What does job security even mean in this new paradigm? How is my daily performance being measured? And if I use this tool to make the company an extra million dollars, how are those financial gains going to be shared? Right. And when a company just pushes an overnight update that installs a new algorithmic tracker that suddenly measures how many keystrokes you make or how long it takes to resolve an email without warning you, that is the breach.

[00:17:17] The trust evaporates instantly. And we have decades of data showing that the consequences of that breach are severe. What did they find? They looked at a vast array of organizations and found that when this unwritten psychological contract is breached, the fallout isn't just a temporary dip in morale. The data shows deep, measurable, long-term declines in employee trust, job satisfaction, and organizational commitment.

[00:17:46] So they don't just get annoyed. No. They fundamentally withdraw their emotional labor. They stop caring about the company's long-term success and turnovers skyrockets. Okay, let's explore that fallout. Because the unwritten contract is breached, the trust is completely gone, the employee firmly believes that the company's true AI values are just, you know, cut costs and squeeze the workforce. Right. What happens next?

[00:18:10] Because from the outside, from the CEO's perspective, it might just look like a few disgruntled employees complaining on private Slack channels. But the research makes it terrifyingly clear that the consequences go way beyond angry workers. It really does. This is where the internal culture problem mutates into a hard operational crisis. In the field, we refer to this as the adoption quality problem. How bad is it?

[00:18:33] To grasp the scale, we need to look at a global study conducted jointly by KPMG and the University of Melbourne. They ran this between late 2024 and early 2025. And the scale is vital to note. They surveyed 48,000 people across 47 different countries. Wait, 48,000?

[00:18:54] So if they're surveying that many professionals about their potentially unauthorized use of AI, I have to imagine the numbers are incredibly skewed. I mean, people generally want to look competent and compliant on surveys, right? You would think so. Are you telling me that a significant portion of these folks actually admitted to bad behavior on a mass scale? They did. Which suggests the actual reality on the ground is likely even worse due to self-reporting bias.

[00:19:21] What KPMG and the University of Melbourne uncovered is an absolute crisis of what we call shadow AI. Shadow AI. Like shadow IT. Exactly. Shadow IT is all your existence. Employees using unapproved messaging tools because the corporate ones are too clunky. But shadow AI is exponentially more dangerous because it actually generates content and makes decisions. What were the stats? The statistics are alarming.

[00:19:49] 66% of employees reported that they routinely rely on AI output without actually evaluating its accuracy. Let me stop you there. Two-thirds of the workforce is taking machine-generated text or code and just passing it along without checking if it's actually true. Yes. The downstream effect of that is captured in the next statistic. 56% of respondents admitted that they have made mistakes in their professional work specifically because they relied on unchecked AI. That is wild.

[00:20:16] But the most disturbing finding, the one that ties directly back to our psychological contract discussion, is that 57% of employees said they actively hide their use of AI from their employer. And they purposely present the AI-generated work as their own original effort. Nearly 60% of professionals are secretly using the tools and claiming the work as their own. It is exactly like doing your high school math homework with a calculator hidden under the desk while the teacher walks by. It really is. But we aren't talking about algebra homework.

[00:20:44] We're talking about professional adults in the knowledge economy, building corporate strategy and writing production code. Why on earth are adults hiding their efficiency? And more practically, how are they even getting away with it? Doesn't the IT department track everything? Don't they see the API calls? Let's define what an API call is for a second, just to ground the technical reality. An API call is essentially the digital footprint left behind every time an employee's software asks the external AI server to do something. Okay.

[00:21:13] So yes, if an employee is using the approved enterprise-provided AI tool, IT can absolutely track their usage. But shadow AI often means the employee is using their personal phone or a personal browser tab, completely disconnected from the corporate network, to run the prompts. Oh, wow. They get the output, and then they manually type it back into the secure corporate system. They go to great lengths to hide the digital trail. Which takes a lot of effort. Why go through all that trouble to conceal it?

[00:21:41] If you view it through the lens of survival psychology, their behavior is entirely rational. Let's go back to the environment of a breached psychological contract. If you firmly believe that every single efficiency you find is going to be weaponized against you to engineer your job away... Concealment is your only defense mechanism. Exactly. Think about the incentive structure.

[00:22:03] If you figure out a way to do a five-hour task in 20 minutes using a clever AI prompt, and you transparently share that with a manager who lacks empathy and just wants to cut costs, what is the guaranteed outcome? The manager says, fantastic work. Right. Since you have so much free time, here are 14 more identical tasks for you to complete by 5 p.m. Oh, and by the way, we are letting your junior colleague go because we clearly don't need the headcount. Right. You are punished for your own innovation. The system literally punishes honesty.

[00:22:33] Yes. So a rational actor simply stops being honest. You do the task in 20 minutes using your hidden AI tool. You spend the next four hours watching YouTube, reading the news, or applying for other jobs. And then at the four-and-a-half-hour mark, you submit the work so it looks like it took the expected amount of time. You systematically conceal the innovation from the organization. Furthermore, if you're actively hiding your use of the tool, you are definitely not going to raise your hand to report bugs when the tool hallucinates.

[00:23:03] Right. You can't admit you were using it. Exactly. You aren't going to participate in focus groups to help the company redesign workflows either. Which means the company is paying an incredibly steep double penalty for their poor AI values. First, they're alienating and eventually losing their top talent because of the breach of trust. And second, they are permanently baking sloppy, unverified, hallucinatory AI work into their core business processes because everyone is too terrified to admit they're using the tools.

[00:23:33] It is an operational nightmare. Generative AI only produces real, sustainable enterprise value when employees engage with it thoughtfully and collaboratively. They need to check its outputs rigorously. Yes. They need to share what prompt engineering strategies yield the best results for specific clients. And they need to collaboratively redesign their own work processes. That level of collaboration requires a foundation of immense psychological safety.

[00:23:59] And a culture of fear produces shadow AI, which eventually produces critical failures. Exactly. But Dr. Westover's analysis points out that this isn't just an internal HR headache anymore. This opacity, the hiding of how AI is being used by both employees and management, is rapidly hitting serious legal tripwires. Okay, let's get into the legal side. How is this escalating? This is a crucial evolution in the landscape.

[00:24:26] Transparency is no longer just a nice-to-have cultural initiative promoted by HR. It is rapidly becoming legally mandated by the state. Really? Where? The analysis highlights a very specific sweeping amendment to the Illinois Human Rights Act, which officially takes effect on January 1, 2026. What are the actual mechanics of that law? Like, how does it change the day-to-day?

[00:24:48] The amendment explicitly requires employers to proactively notify workers and prospective candidates about the use of AI in any employment decisions. So if a company is using an algorithm to screen resumes or optimize shift scheduling or evaluate performance metrics... Or to make disciplinary and termination decisions, yes. The company is now legally required to disclose that fact to the affected individual.

[00:25:14] You can no longer hide the AI behind the curtain of management discretion. That's massive. And furthermore, and this is the truly massive legal shift, the law treats algorithmic discrimination as a definitive civil rights violation. Let's define algorithmic discrimination in this context, because I think people assume AI is naturally objective, right? It's just math. How does an algorithm discriminate?

[00:25:38] Algorithms discriminate because they are trained on historical human data, which is inherently biased. And they often rely on what we call proxies. What's a proxy? A proxy is a seemingly neutral data point that actually correlates heavily with a protected class. A classic example is a zip code. An AI might look at historical hiring data and determine that people from a specific zip code tend to leave the company faster. So it starts automatically filtering out resumes from that zip code.

[00:26:06] Exactly. The AI doesn't know what's being racist or glasses. It just sees a statistical correlation. But in American cities, zip codes are often highly correlated with race and socioeconomic status due to decades of systemic redlining. Ah, I see. So the AI ends up illegally discriminating based on race using the zip code as a proxy. Wow.

[00:26:27] So under this new Illinois law, if a company's internal enacted value is to use AI secretly to manage the workforce and cut corners on HR, they aren't just going to lose employee trust. They are going to get sued by the state for civil rights violations. They absolutely are. And presumably if Illinois is doing this, California, New York and other states are immediately going to follow suit. That is the expectation.

[00:26:51] Organizations that have relied on opaque black box AI deployments to avoid having difficult conversations with their workforce are going to find themselves completely exposed. They will have to explain their choices to state regulators first and only then to their angry employees. So the legal exposure forces the entire values test out of the private boardroom and into the public square. Yes. So the anxiety permeating the workforce isn't just about losing a paycheck and the corporate risk isn't just about getting sued.

[00:27:19] Because we're talking about generative AI, which specifically targets professional work, knowledge work and creative work, the threat strikes at something much deeper. It threatens professional identity. This is a vital pivot in understanding the true depth of the AI values test. Let's unpack that.

[00:27:36] When we look back at previous waves of industrial automation, like the assembly line or robotic arms and auto manufacturing or automated toll booths, we were largely talking about the replacement of routine physical manual labor. Right. The physical toll on the human body was alleviated, even if the economic disruption was severe. Exactly. But generative AI is fundamentally different. It can write complex Python code. It can draft nuanced legal briefs. It can generate compelling marketing copy.

[00:28:06] It can create intricate illustrations. It's coming for the exact skills that people spent four years in college or a decade in the field trying to master. It's coming for the cognitive output that people base their entire self-worth and professional pride on. And because of that profound shift, people are experiencing an existential dread. They feel that their deeply hard-won expertise is being commodified, treated as nothing more than raw material to feed a machine. It's a horrible feeling.

[00:28:34] Imagine spending 20 years perfecting your craft as a technical writer. You know exactly how to explain complex engineering concepts simply. Then, your employer feeds all of your past manuals, your entire life's work, into a proprietary large language model. Oh, man. Now, a junior employee with zero technical writing experience can generate your writing style with the click of a button. Economically, you might still have a job, for now.

[00:29:01] But psychologically, the employer has stripped you of your unique professional identity. You are no longer a craftsman. You are training data. Which brings us to one of the most visible high-stakes battles over this exact issue, the Writers Guild of America strikes. Dr. Westover's analysis does a deep dive into the WGA strikes of 2023 and the subsequent updated agreement in 2026. It's a fascinating case study.

[00:29:26] And I have to say, looking at the mechanics of what the writers actually fought for, this is where the conversation gets incredibly interesting to me. Because the prevailing narrative in the media was that the WGA was trying to smash the machines and ban AI entirely. But that wasn't their strategy at all. No, and that is a very common, very reductionist misconception. The dispute was not about pretending the technology didn't exist. It was fundamentally a battle over algorithmic justice, economic structure, and identity protection.

[00:29:55] Important to note, we are just looking at this impartially as a case study. Exactly. From a structural perspective, the writers recognized that generative AI was a permanent reality. But they wanted to forcefully and legally define the terms of its entry into their craft. They wanted to draw a very clear boundary between human creative labor and machine generation. Let's look closely at the specific protections they secured because they function as a brilliant blueprint for how other knowledge industries might handle this.

[00:30:23] First, they established that AI cannot write or rewrite literary material. Which means an AI program can never be credited as a writer on a project. Right. Second, and this seems like the linchpin, AI-generated material does not count as source material. I want to pause here because to an outsider, arguing over the phrase source material sounds like trivial legal semantics. Why did that specific phrase matter so much in Hollywood?

[00:30:51] It matters because the definition of source material is the entire economic engine of a writer's career. In the Hollywood guild system, whoever creates the original source material holds the underlying rights to the characters, the world-wielding, and the narrative structure. Ah, so they get the primary credit. They get the primary credit. They get the largest upfront fee. And crucially, they get the residual payouts every time that show or movie is re-aired or streamed.

[00:31:18] Okay, so if the studios were allowed to use an AI to generate a terrible garbled 90-page script and then hand that AI script to a human writer and say, here is the source material, we just need you to polish it, the human is economically reclassified. Exactly. They are no longer the creator. They are legally just a punch-up editor, getting paid a fraction of the original fee and losing out on the long-term residuals.

[00:31:44] By explicitly stating AI cannot create source material, the WGA protected the economic foundation of human origination. That is brilliant foresight. They protected the credit, which protects the money. What else did they secure? Third, they established that companies cannot force writers to use AI software if the writer prefers their traditional process. And finally, in the updated 2026 agreement, they secured strict notice requirements.

[00:32:13] Meaning, if a studio wants to take existing scripts out of the vault to license them for training new AI models, they legally have to notify the writers. Yes. If we look at the situation purely as an analytical case study in labor dynamics, setting aside any political viewpoints on unions or studio management, what the WGA achieved was the successful codification of an AI values test. They essentially demanded that the studios make their enacted values explicit in a legally binding contract.

[00:32:43] Exactly. The employees cared deeply about the terms on which this tech entered their daily workflow. It wasn't just a battle over compensation. It was a battle over meaning and professional dignity. They collectively refused to be reduced to mere editors of machine output. They refused to be, as the original text put it, editors of machine vomit. That is a very visceral way to put it. But honestly, that feeling is permeating the entire knowledge economy right now.

[00:33:10] So many white-collar workers feel like they are being quietly demoted from creators to janitors. Janitors spending their days cleaning up the hallucinatory mess the AI leaves behind. Yes. And that feeling of demotion, that stripping of professional dignity, is the absolute quickest way for an organization to fail the values test and trigger an exodus of their most talented people. Okay. So we have firmly established the intensely high stakes of this test.

[00:33:36] We've looked at the entire spectrum of failure, from employees secretly hiding their calculator homework under their desks, all the way up to industry-wide billion-dollar strikes. It's a broad spectrum. So the obvious pressing question for anyone trying to run a business right now is, how are companies actually navigating this in the real world? The analysis provides some incredible detailed case studies of the AI values test in action. It does. Let's look at the playbook. The good, the bad, and the brutally honest.

[00:34:05] The first major lesson that emerges from analyzing these corporate case studies is what we might call the context versus candor problem. The core unbreakable principle here is that an organization must thoroughly explain its overall strategy before they send out a tactical memo. Because if you announce a tactic without the strategy, disaster follows. Complete disaster. And the poster child for getting this horribly wrong, at least in their initial rollout, seems to be Duolingo. What exactly happened there?

[00:34:32] In April 2025, Duolingo's CEO, Louis von Ahn, sent out a company-wide internal memo. He announced that the company was aggressively pivoting to become AI first. Okay. Pretty standard corporate speak. Right. The memo detailed immediate plans to gradually stop using human contractors for translation and content generation work that the new AI systems could now adequately handle. Still pretty standard.

[00:34:58] But the most alarming part of the memo was a mandate stating that the company would only approve new headcount for teams if those teams could definitively prove they couldn't automate more of their existing workload using AI. Oof. Let's put ourselves in the shoes of a Duolingo employee reading that on a Tuesday morning. My blood pressure is spiking through the roof, right?

[00:35:19] The CEO is literally saying we are actively replacing human labor with AI and we aren't hiring anyone new to help you unless you prove to us that the robots are incapable of doing your job. Yeah. It feels incredibly adversarial. It is entirely adversarial. And predictably, there was massive internal panic and a very public backlash. You know, the company became basically a cautionary tale overnight. Yeah, I bet. Yeah. What made that? What was the... Yeah, yeah.

[00:35:48] The critical failure was that the memo lacked any reassuring broader context. You know, it wasn't until four months later in August of 2025 that von Ahn was forced to do kind of a major clarifying interview with Fortune magazine to stop the bleeding. Four months is a really long time in the corporate world. A very long time. Yeah. He had to publicly admit that the initial memo severely lacked context. And he explicitly clarified that the company had never laid off any full-time employees due to AI and had absolutely no plans to do so.

[00:36:18] So they were just changing the nature of the roles they were recruiting for. Exactly. He explained they were still hiring at the exact same overall pace. But the damage to the psychological contract was already done. Think about those four months. For four months, those full-time employees were living in a state of constant, low-grade terror. They were hyper-anxious. They were probably updating their resumes. They were taking calls from recruiters. And their productivity likely tanked.

[00:36:46] All because the tactical mandate was announced without the strategic context. The lesson is brutally clear. If leadership announces a shift with gaps in the narrative, the workforce will automatically, instinctively fill those gaps with the most threatening, catastrophic interpretation possible. Precisely. Now, if we look for a contrasting approach, we can examine Shopify. In April 2025, Shopify CEO Toby Lutka took a very different path. What did he do? He didn't just announce AI as a vague corporate aspiration.

[00:37:16] He made proficient AI usage a mandatory baseline expectation for employment. He tied it directly and explicitly to employee performance reviews. Oh, wow. He essentially communicated, using these tools to augment your output is now a core part of your job description. And your continued employment and compensation will be evaluated based on how will you do it. That is definitely clear. There is absolutely no ambiguity there, which solves the context problem.

[00:37:44] But practically speaking, how does a manager actually grade that? That's the tricky part. Is the manager just checking a dashboard to see if I logged into the enterprise AI software three times this week? Because if that's the metric, I'll just write a script to log me in automatically. And that is the inherent danger of tying AI to performance metrics poorly. If you just measure superficial usage or login rates, you get garbage compliance. Evaluating AI use meaningfully means looking at output speed and quality relative to peers.

[00:38:13] A good manager has to ask, are you just using the AI to do more of the exact same routine tasks? Or are you leveraging it to unlock completely new higher level capabilities that benefit the team? And I guess that depends on the support they get. Exactly. As Dr. Westover notes, whether employees view Shopify's mandate as an empowering challenge to upskill or a terrifying threat to squeeze more work out of an exhaustive workforce depends entirely on the support structure.

[00:38:39] If Shopify provides massive paid training during work hours, it's a win. If they just demand higher output without support, it will completely backfire. And then there is the Amazon approach. In June 2025, Amazon CEO Andy Jassy sent out a memo that was just brutally, unflinchingly candid. Very candid. He basically told the workforce, look, as we integrate generative AI, it is going to reduce the corporate workforce over the next few years. There will simply be fewer people doing certain administrative and coding jobs.

[00:39:09] That is the ultimate theft of extreme candor. Jassy didn't attempt to sugarcoat the reality or hide behind vague PR speak about synergies. Right. He just laid it out. But the analysis warns that while this level of candor is ethically superior to vagueness, delivering candor about impending job cuts without making equally specific concrete commitments about transition support, severance packages, or internal retraining. That's going to trigger that massive wave of self-preservation and shadow AI we discussed earlier. Immediately. Yes.

[00:39:40] Which perfectly sets up the next category of case studies. How organizations are moving from unwritten, implicit psychological contracts to actual, explicitly written AI employment deals. And the gold standard for this, according to the research, is IKEA, specifically the Inca group. This story is fascinating because it defies every corporate stereotype. What did they do? The IKEA case study is truly the textbook example of a redeployment-first enacted value.

[00:40:09] The timeline here is important. Starting back in 2021, IKEA realized that their AI customer service bot, which they named Billy, was becoming incredibly proficient. Over a two-year period, as Billy learned from vast amounts of data, the bot successfully took over 47% of all routine customer queries to their global call centers. Almost half of the entire human workload just vanished into the algorithm.

[00:40:35] For 99% of companies, that is the exact moment the CFO rubs their hands together, calls a board meeting, and says, Fantastic news. We can immediately fire 47% of the call center staff. Our quarterly margins are going to look incredible. But IKEA deliberately chose a different path. Instead of capturing those gains through mass layoffs, they took 8,500 of those call center workers and aggressively, systematically retrained them to become remote interior design advisors. Wow.

[00:41:05] They took the massive amount of human hours freed up by the AI and purposefully reinvested it into creating a completely new, higher-value, personalized human service. The mechanics of that transition must have been incredibly difficult. I mean, how do you take someone whose job was answering phones about missing screws and turn them into an interior designer? It's not easy. It requires a completely different skill set, spatial awareness, deep product knowledge, high-level empathy, and complex problem solving.

[00:41:32] It required a massive financial investment in training and a deep belief that their employees were capable of high-level advisory work if given the right tools. And the critical question is, did that investment pay off? Right. Right. From a cold, hard business perspective, did this feel-good strategy actually make money? It was wildly empirically successful.

[00:41:53] In the 2022 financial year alone, that newly created remote interior design service generated 1.3 billion euros in revenue. A billion euros. A billion euros. Yes. Yes. That accounted for 3.3 percent of their total global sales, and the company is now aiming to push that to 10 percent by 2028. And when an IKEA executive was subsequently asked by the press if AI would reduce their future headcount, they honestly answered that they weren't seeing that trend.

[00:42:22] Because every time AI freed up time, they were constantly finding new profitable problems for their human workforce to solve. Exactly. That is incredible. They didn't just issue a stated value memo telling their employees, we value your development. They enacted that value to the tune of 8,500 retrained careers and over a billion euros in new revenue. It's a powerful example. But let me push back a little here and ask a pointed question about the reality of retraining.

[00:42:51] Let's look at the Accenture case study from September 2025. CEO Julie Sweet announced a massive $865 million corporate restructuring, and she was extremely explicit. She stated publicly that the company would exit staff who simply could not be retrained or upskilled fast enough for the new AI-driven landscape. She did.

[00:43:10] But in the exact same breath, she highlighted that the company had successfully created 77,000 newly skilled AI professionals internally through their training programs. So my question is, does an approach like Accenture's, which is very stark, very adapt to the new tools or you are fired, but clearly stated up front, does that actually build more trust than a company that makes sweeping, highly unrealistic promises like, don't worry, no one will ever lose their job to AI?

[00:43:40] The decades of research on psychological contracts are unequivocal on this specific point. Yes. The stark but stated approach builds exponentially more trust. Really? Yes. Bounded, realistic promises that are strictly kept are infinitely better for organizational health than generous utopian promises that are eventually broken. Employees are functioning adults. They understand the macroeconomic reality that businesses have to adapt to survive. They just don't want to be lied to.

[00:44:08] What they absolutely cannot tolerate is being patronized or lied to. When a CEO stands up and says, no one will lose their job in the midst of a technological revolution, every single employee in the room knows that is highly probable to be a lie. And they immediately, permanently lose respect for the leadership. So it's about defining the boundaries of the relationship clearly. Accenture essentially said, we will spend hundreds of millions of dollars to try and build a bridge for you to cross this chasm.

[00:44:38] But if you refuse to walk across it or you simply can't make it, you are out. It's tough. It's uncompromising. But it is an honest, explicit deal. Exactly. It takes the terrifying ambiguity of the implicit psychological contract and makes it explicitly clear, allowing the employee to make an informed choice about their own career. OK, so let's say an organization has done the hard foundational work. They've announced the plan with proper context.

[00:45:05] They've written an explicit, honest deal with the workforce. But how do they actually govern this machinery day to day? That's the next big hurdle. Because generated AI isn't static. It learns. It morphs. It hallucinates. And it occasionally breaks. And more importantly, once this massive machine is running and actually generating profound productivity gains, who gets the spoils? This brings us to the critical issue of sharing the gains and protecting the quality of the output.

[00:45:34] The specific way an organization chooses to distribute the windfall of AI productivity is the ultimate undeniable proof of their AI values. And there is a phenomenal, highly publicized cautionary tale in Dr. Westover's analysis about exactly this issue, the Klarna U-turn. Let's deep dive into the mechanics of this one because it is an absolute roller coaster of corporate hubris.

[00:45:57] In February 2024, Klarna, the massive global buy-now, pay-later fintech giant, proudly announced to the world that their newly deployed AI assistant was doing the equivalent work of 700 full-time human customer service agents. 700? Yes. They boasted that the AI was independently handling 75% of all customer chats. On the back of this apparent monumental triumph, they instituted a total global hiring freeze and began actively reducing their human headcount through natural attrition.

[00:46:26] It was heralded across the business press as the ultimate AI success story in corporate efficiency. I remember seeing those headlines everywhere. Yep. Every CEO on LinkedIn was pointing to Klarna and posting things like, see, this is the future. We need to implement this immediately. But what actually happened next? The harsh reality of the human customer experience caught up with the technological hype.

[00:46:48] By May 2025, just over a year later, the CEO, Sebastian Siamakowski, had to execute a humiliating, complete reversal, of course. A complete reversal? Yes. He gave an interview to Bloomberg, admitting that the company was actually aggressively hiring human customer service agents again. Why? Did the AI system technically break down under the load? The AI didn't break technically. It was still generating text at lightning speed.

[00:47:17] But it broke functionally in terms of client satisfaction. Siamakowski had to admit that the purely AI-driven approach produced significantly lower quality service. Ah, so the customers hated it. Customers were incredibly frustrated. While the AI was fantastic at answering basic questions like, what is my balance? It completely failed when issues became complex. It lacked the necessary nuance, the human empathy required when a customer was in financial distress,

[00:47:45] and the creative problem-solving capability of a seasoned human agent navigating edge cases. So they learned the hard way. Klarna painfully realized that protecting a robust human option was absolutely essential for maintaining service quality and customer trust. They had to publicly backtrack, admit the error, and heavily reinvest in human support. When you look at the timeline of that failure, is it possible we are just living through a classic AI hype cycle

[00:48:13] where executives were so temporarily blinded by the sheer intoxicating potential for immediate cost-cutting that they entirely forgot they actually have human customers who actively hate talking to unhelpful chatbots? That is a highly accurate assessment of the current cycle. And it highlights the immense systemic danger of what the analysis calls the gains-capture approach.

[00:48:37] When an organization views generative AI solely as a mechanism to capture financial gains through aggressive headcount reduction, they trigger a cascade of negative effects that eventually degrade the core product. What does that cascade look like internally before it hits the customer? Let's look at the mechanism of productivity. A major study by Brian Olson and colleagues in 2025 found that in customer support settings, AI assistants created an average measurable productivity boost of 14%.

[00:49:06] 14% is a solid bump. And it notably helped the less experienced workers the most by giving them access to best practices. Now, if corporate leadership takes that 14% productivity boost and immediately translates it into firing 14% of the workforce to capture the margin, what powerful lesson does that teach the remaining 86% of the employees? It teaches them that every single time they figure out a way to use the AI tool more efficiently, they are literally tightening the noose around their own necks, or the necks of their friends sitting in the next cubicle. Exactly.

[00:49:36] And if you look at the ripple effect of that realization, it instantly suppresses all future experimentation. If innovation equals termination, employees will stop innovating immediately. And the shadow AI problem returns in full force. Yes, and the company's productivity flatlines. To actually pass the AI values test, organizations must reinvest a visible, meaningful share of those time savings back into the workforce. That could mean creating entirely new, higher value roles like IKEA did.

[00:50:04] Or reducing workloads to prevent burnout, or giving them paid time for professional development. Exactly. So you have to systematically share the wealth with the people generating it, or the wealth generation engine shuts itself down in self-defense. But beyond just sharing the financial gains, the analysis also emphasizes the importance of governing visibly, and giving workers a literal, structural voice in the process. Yes. Visible governance is the only way to sustain trust over time.

[00:50:32] The article highlights how some major tech companies are trying to build this architecture right now. What's a good example of that? Well, in December 2023, Microsoft formed a very public, highly structured partnership with the AFL-CIO, which is the largest federation of unions in the United States. And again, keeping our neutral analytical hats on regarding labor politics, what was the actual structural purpose of that specific partnership from an AI values perspective? What were they trying to achieve?

[00:51:00] The stated contractual goals of the partnership were to formally incorporate worker perspectives and frontline operational expertise directly into the development cycle of Microsoft's AI technology. It also included a neutrality framework for union organizing. From a signaling perspective, Microsoft was demonstrating a willingness to give organized labor a formal structured channel

[00:51:25] to voice legitimate concerns about how the technology is built and deployed in the workplace, rather than just forcing the finished product on them post-development and demanding compliance. And Salesforce has been pioneering a version of this for quite a while, haven't they? Yes. Salesforce is frequently cited as an early pioneer in visible governance. Back in 2018, long before the current generative AI boom, they created a dedicated Office of Ethical and Humane Use of Technology. That was early. It was.

[00:51:55] They established internal advisory councils that intentionally included frontline employees, not just C-suite executives. They published clear, trusted AI principles. And crucially, this is the vital mechanism, they tied those principles to actual mandatory review processes within the product development cycle. That is the absolute key right there. Tied to actual review processes. Because I feel like every mid-sized company in the world currently has a glossy PDF somewhere on their internet

[00:52:23] that says, our AI principles, be fair, be nice, protect privacy, don't be evil. No, definitely. But if there isn't a hard structural mechanism to actually hit the brakes on a lucrative project when it violates those principles, the PDF is worthless. It is actually worse than worthless. It functions as a highly hypocritical signal. If employees see a massive glaring gap between the noble stated principles in the PDF and the ruthless enacted reality of the software deployment, they will assume the gap is deliberate and cynical.

[00:52:51] So visible governance is essential. Visible governance, having a diverse council, having an anonymous hotline, having a structured appeals process if an algorithm unfairly flags or fires you, is the only way an organization proves that their principles actually have teeth. Okay, so we have discussed the overarching strategy, the explicit deal, and the corporate governance structure. But let's get incredibly real for a second about how a company actually functions. The CEO writes the grand strategy memo.

[00:53:20] The HR department drafts the ethical principles. The legal team checks the compliance. But the average employee sitting at their desk on a Tuesday afternoon doesn't interact with the CEO or the VP of HR or the general counsel. They interact almost exclusively with their direct manager. And this brings us to the final and perhaps most operationally crucial section of the entire analysis, building long-term capabilities. An organization can draft the most enlightened progressive AI memo in the history of capitalism,

[00:53:49] but passing the AI values test day after day, month after month, falls squarely and heavily on the shoulders of one specific group, middle managers. Because the middle manager is the actual delivery mechanism for the corporate culture. Exactly. Most employees experience their corporation's AI values exclusively through the lens of their direct manager. Think about the daily micro-interactions that define a job. Like what?

[00:54:16] It is the manager who decides if an employee using ChatGPT to outline a quarterly report is considered taking brilliant initiative or lazy cheating. It is the manager who handles the delicate situation when an honest employee admits that an AI hallucination caused a minor statistical error in a client presentation. Right. If the manager freaks out, yells at the employee, and formally punishes them for the AI-assisted error, that employee learns a permanent lesson.

[00:54:45] Never, ever admit an AI mistake again. Shadow AI is born in that single, poorly handled interaction. Shadow AI is born in that single, poorly handled interaction. Right now, organizations are communicating these grand, sweeping AI visions from the top floor, but they are leaving their middle managers completely unequipped to handle the messy, psychological human realities unfolding on the ground floor. So what do managers need?

[00:55:07] The analysis stresses that managers desperately need three specific capabilities to survive this transition without destroying their teams. First, they need actual foundational AI literacy. Which means they need to actually know how the underlying technology works, not just what the vendor's sales brochure promises. Correct. They need to understand what large language models can and cannot reliably do, where they are statistically likely to fail or hallucinate,

[00:55:34] and how to properly evaluate AI-assisted work versus purely human work. And are they getting that training? The data shows that globally, only 47% of employees have received any formal AI training. This means that millions of middle managers are currently trying to lead their teams through the greatest technological revolution since the Internet, and they fundamentally do not understand the technology themselves. That is a terrifying recipe for disaster. If the manager doesn't know how the API works, how can they guide the team?

[00:56:04] What is the second capability managers need? They need highly developed conversation skills. Understanding the technology is useless if you don't know how to talk to a terrified human being. Managers must have the ability to sit down with an anxious, defensive employee and discuss job insecurity, impending role changes, and fairness concerns with deep honesty and empathy. That's a tough conversation to have. It is. The decades of research on organizational justice are incredibly clear on this point.

[00:56:31] Even if a management decision is difficult or negative, like a specific role being eliminated due to automation, if the explanation provided is thorough, transparent, and the interpersonal treatment is highly respectful, employees are far more accepting of the outcome. So they have the tech literacy and they have the empathy to have the hard conversation. What is the third capability? Decision rights. Managers need clear, defined boundaries, but within those boundaries, they need the discretion and the actual authority to act.

[00:57:01] They need to know they have the corporate backing to reward an employee who finds a great AI shortcut, rather than feeling obligated by a rigid corporate policy to immediately report that efficiency to HR so the time savings can be harvested for the bottom line. Well, that makes sense. If managers aren't structurally equipped with these three things, literacy, conversation skills, and decision rights, then the company's AI values will be wildly, destructively inconsistent.

[00:57:30] An employee's entire experience of the technological transition will depend entirely on a roll of the dice. Did they draw a good, literate manager or a paranoid, unequipped one? So if you are an executive listening to this, how do you actually know if your managers are pulling this off? How do you measure if the organization is passing the test? Because you can't just walk around asking people, hey, do you trust me? No, you can't. You have to systematically measure trust, not just software adoption rates.

[00:58:00] Right now, most companies have elaborate dashboards, tracking how many API calls their teams make, or how many employees logged into the enterprise AI portal this week. But adoption is not a proxy for trust. Because a terrified employee who is hiding their real work and just clicking buttons to meet a quota generates adoption. Exactly. So what specific metrics should leadership be looking at? The Gallup data gives us a massive clue here.

[00:58:26] They found that an employee's self-reported understanding of the AI plan is a massive leading indicator of their overall comfort, preparedness, and retention. Oh, interesting. So companies need to run regular anonymous pulse surveys, asking very pointed, specific questions. Do you clearly understand our strategic AI roadmap? Do you feel psychologically safe disclosing your AI use to your direct manager? Do you believe the AI-driven decisions affecting your daily work are fair?

[00:58:53] Do you see the productivity gains being shared equitably across the team? You essentially have to audit the vibes. You have to put hard data behind the unwritten psychological contract. And you have to permanently bake this explicit deal into your employer value proposition, your EVP. This is how you pitch your company to new recruits in the market. Why is that so important for recruiting? According to Benjamin Schneider's famous 1987 attraction, selection, attrition model,

[00:59:20] organizations naturally become more homogenous over time because people who fit the culture stay and those who clash eventually leave. If you integrate your explicit, verifiable AI commitments into your recruitment pitch, if you tell them exactly what the enacted values are on day one, you will attract candidates who align with that specific culture and you will repel the ones who don't. We have covered a truly massive amount of ground today.

[00:59:47] We've explored the deep paradox of workers wanting the technology but fearing the deployment. We've unpacked Spence's signaling theory, the psychological survival mechanism behind shadow AI, the economic identity battle of the Writers Guild strikes, the Klarna customer service disaster, and the incredibly heavy burden placed on middle managers. It's a lot to take in. It is. If I had to synthesize all of this into a single thought,

[01:00:12] it seems like the AI values test isn't a one-time pass-fail exam you cram for during a quarterly retreat. It is a continuous, relentless, daily assessment by your entire workforce. Yeah. I mean, that's the perfect sort of summary of the landscape. And the core, you know, the inescapable takeaway from Dr. Westover's research is this,

[01:00:36] that the specific way a company chooses to adopt and deploy AI is the message. I mean, that is, you know, what was it? Marshall McLuhan. Yeah. The medium is the message. Yes. That is what he was talking about. Yeah. You cannot separate the capability of the technology that you're using it from the sociology of how it's deployed. That's if the glossy, optimistic story told in the, you know, the boardroom does not perfectly match the, you know,

[01:01:04] the gritty operational reality of, you know, the Tuesday morning team meeting, the talent is going to, they're going to simply leave. And they will take their hard-won skills, their institutional knowledge, their willingness to innovate, like you said, straight to a competitor who actually knows how to pass the test. Absolutely. Thank you all for joining us on this incredibly deep dive. It is abundantly clear that artificial intelligence is no longer just a cool productivity hack.

[01:01:30] It is a harsh, unforgiving mirror reflecting an employer's true priorities. It absolutely is. And I would like to leave everyone listening with a final forward-looking thought, something to chew on that builds on everything that we've explored today about alignment and values. Let's hear it. As AI agents become increasingly autonomous, as they evolve from simple chatbots into deep decision-making components of our daily corporate workflows, it raises a fascinating question about the near future of the labor market.

[01:02:00] In five years, when you sit down for a high-stakes job interview, will you be interviewing a company's AI agent to decide if you want to work for them? Or will the AI be interviewing you, rigorously testing your personal values to see if you are a fit for the machine's culture? Wow. That is slightly terrifying. But given the trajectory we are on, absolutely possible. Something to deeply think about the next time you log in.

[01:02:30] Thank you for listening, and as always, keep questioning the world around you. Thank you.すごistic完成 Thank you. Thank Thank you.