#27 Human-centric Indispensability: How will humans remain relevant in the future job market?

Foreign. Welcome to my podcast, Future of Work, Future Skills and AI. I'm Yasmin Weiss, professor of Artificial Intelligence and the Future of Work. This is no ordinary podcast I created using the AI tool NotebookLM. The characters you hear are two AI voices. However, the content comes from me. Yasmine, each week you will learn how to work better and smarter with AI. Right now, major tech companies are laying off thousands of human workers. And it's not because they're losing money, right? Not at all, no. It's actually to free up cash to buy the AI servers that are replacing those exact workers. So welcome to the Deep Dive. Glad to be here. Today, our mission is to explore a highly relevant, maybe a little intimidating, but absolutely vital framework published on LinkedIn by Professor Dr. Yasmin Weiss. And she is definitely one of the leading voices on the future of work with AI right now. Exactly. So our goal today is to really unpack how you, the listener, can remain strategically indispensable, how you can maintain control in a job market that is just rapidly shifting toward what she calls coopetition, which is basically the simultaneous cooperation and competition between humans and machines. Yeah, and it's a fundamental rewiring of the labor market, isn't it? Oh, absolutely. I mean, cognitive work, the deep analysis, the synthesis, is being aggressively consumed by agentic AI. And you know, we can't ignore the physical side either. Physical AI isn't just like robotic arms on a pre programmed assembly line anymore. It's about machines that possess actual spatial intelligence. They can adapt to unstructured environments in real time. So we're really seeing algorithms encroach on both ends of the human capability spectrum. Okay, let's unpack this. Because to really understand this coopetition, we have to look at two massive concepts from her research. The acronyms. Right, the acronyms, they sound like buzzword, I know, but they represent very real, very harsh economic realities. I'm talking about fobo, the fear of becoming obsolete, and Foley, the fear of losing agency. Yeah, and minimizing these fears isn't about, you know, just ignoring them or pretending the tech isn't there. Right. You can't just bury your head in the sand. Exactly. It's about finding a proactive strategy. Because agentic AI, unlike traditional software that just runs predefined loops, it's actually given an objective and then determines its own path to achieve it. Which is wild. It is. And that autonomy is what triggers phobo and foley. It's not some irrational anxiety when a system can independently strategize human utility naturally gets called into question. So as we dive into this, I want you to mentally audit your own daily tasks. Like think about what you did yesterday. That's good exercise. Yeah. Like what was just routine execution and what actually required genuine unstructured human intuition. Because to survive this, we have to understand what the actual frontier of human AI collaboration looks like right now. Professor Weitz actually provides a personal case study of this. She does. And her daily workflow is just a perfect illustration of this whole coopetition thing in action. Right. She doesn't wake up to a traditional email inbox anymore. I mean, that's gone. Oh no. She wakes up to what's essentially a multi agent platform. While she sleeps, she has this AI team working for her. Yeah, Agent One. Exactly. Agent One is her morning briefing agent. It synthesizes all this global Data that Agent 2, or Research Agent, collected overnight on the future of work. But the mechanics of the next step are what's really crucial here. Yes. Agent 3. Right. Agent 3 is the fact checker. It doesn't just passively read the summary. It actively interrogates the output from those first two agents. It highlights contradictions, checks the logic, and then gives her strategic recommendations. It's basically like managing a team of incredibly eager, totally tireless interns. Right. Pretty much like they work the night shift and just leave a flawless executive summary on your desk by dawn. This is a great analogy, but there's a critical nuance here from the source. She is actually spending what she calls a nerd day at Microsoft right now. A nerd day. I love that. Right. And she's doing this to integrate even more agents, like an agent four for social media, an agent five for writing, and even an overarching chief of staff agent to manage all the sub agents. But here's the key. She retains the final critical assessment, what the source refers to as human refinement. Human refinement, yes. She's an AI augmented professional. Her goal is to reallocate her time to where human intelligence adds unique value. You know, not just automate herself entirely out of the loop. Right, Automating the preparation so she can focus on the actual strategic execution. Exactly. I mean, it sounds amazing in theory, having a personal AI Chief of staff. But this extreme level of automation leads us directly into those two dark questions driving Fobo and Fola. It really does. Because when the digital in interns become vastly faster and more capable than the human manager, the hierarchy just breaks down. So let's look at Fola. The fear of losing agency, the Source asks a very dark Are we racing towards collective suicide? And look, I know that sounds like hyperbole, it really does, but it's not. It is the literal fear of losing control over systems that are becoming increasingly capable, executing decisions way faster than we can even track. Yeah, and the research highlights a specific real world example of this. The Claude Mythos system by Anthropom. Oh, right. Project Glasswing. Yes. So this system spots security flaws, but with just unprecedented precision and speed, it drastically lowers the barrier for complex cyber attacks. Because traditional human driven security updates just. They literally can't keep up. Right, because traditional cybersecurity relies on human engineers finding a flaw, writing a patch and deploying it. And that takes weeks, sometimes months. And Claude Mythos operates at machine speed. Exactly, which is why Anthropic didn't release it publicly. Okay, but that isolated decision brings up the second uncomfortable question, and this is the core of fobo, the fear of becoming obsolete. The question is, are we entering an automation race to the bottom? And if we connect this to the bigger picture, the economic evidence is right there. I mean, look at Meta, Facebook, Instagram, WhatsApp. They are cutting roughly 10% of their global workforce by the end of this month, May 2026. 10% and freezing of open role. Right, and what's Meta's justification for this? They explicitly state they need to fund massive AI infrastructure investments. Billions of dollars in server clusters. Exactly. And those investments are being paid for by the personnel savings and the efficiency gains created by the AI itself. The AI is literally cannibalizing the human workforce to fund its own physical expansion. Okay, but let me push back on this a little bit. We look at the Claude Methos delay, Project glasswing. Yeah. Does that prove that tech companies actually can responsibly self regulate? Or is that just a terrifying warning that we're already totally dependent on the goodwill of a few tech giants? Like, are we just hoping they don't break the world? It's a really valid question. We are largely operating in a regulatory vacuum right now. Right. Anthropic's restraint is great, but it's an anomaly. The relentless pressure of the market dictates that if one company doesn't deploy a capability, a competitor probably will. We can't just wait around. Exactly. Because tech companies are moving at light speed. We can't just wait for them to protect our jobs or, you know, our agency. Which brings us to the cure, because we have to implement Professor Weiss's framework, the human centric indispensability framework. That's a mouthful, but yes, it is. But element one is crucial. It's personal AI resilience. Right. And this is really the foundation. It's a complete mindset shift. You have to view AI as a complementary partner. Like a dynamic dance. Yes, exactly. It's a dance where sometimes the human leads and sometimes the machine leads completely based on who is stronger in that specific context. Okay, so that's the mindset. But here's where it gets really interesting. Element two is what she calls radical adaptability. Radical adaptability? Yeah. And this goes way beyond just the usual corporate like lifelong learning stuff. It requires what she calls iterative identity formation. This is where it gets tough. Very tough. I mean, the detail from the text that stood out to me. She says you have to actively unlearn parts of your professional self image. Like discard your working methods every few months just to make room for new ones. And I have to push back here. Go for it. Unlearning who you are professionally every few months, that just sounds incredibly exhausting. It does. How is that even sustainable for the average worker? Just constantly tearing down your expertise and starting over? It sounds exhausting because it is. But we have to focus on the why. You have to acknowledge that business models today can radically change overnight. That's true. A newly deployed AI agent can render a workflow or even a whole department completely obsolete by Tuesday morning. Wow. Yeah. So past successful behaviors, they mean nothing if the game changes. Radical adaptability is literally the only way to keep pace with the velocity of technological change. Your identity has to be tied to your ability to evolve, not to the specific tasks you do. Okay, I hear that. So mindset and adaptability, that's the internal preparation. But what about the external defenses? Like how do we practically defend against Phobo and Phobos in the real world? That brings us to element three. Strategic skill stacking. Skill stacking. This is the exact formula to combat Phobo. You need to combine at least three non correlating human skills. Three skills? Yes. The source specifies this exact combination. Deep domain knowledge or experience, which, you know, creates human intuition plus strong social skills. And finally the ability to handle contextual ambiguity or physical unpredictability. Contextual ambiguity. So basically knowing what to do when everything suddenly goes off script. Exactly. It's funny, when I read this, I immediately thought of it like creating a password. Oh, I like that. Right, like an impossible to guess password, but made out of human traits. If your career password is just one skill, like, I don't know, data analysis and AI can hack that or automate it in seconds. Absolutely. But if your password combines, say, deep domain intuition plus high EQ to manage a panicked plus comfort and total chaos, it becomes statistically improbable for a machine to replicate you. That is a perfect way to look at it. You are forcing the AI into an out of distribution scenario. Machines fail when they don't have historical data to predict the outcome. But human intuition actually thrives in those novel intersections. Okay, so that solves fobo. We make ourselves irreplaceable. But what about foa? What about the fear of losing agency over the machines themselves? Right, because being irreplaceable doesn't help if the machines running the infrastructure just go completely, completely rogue. Exactly. So to combat foa, we need element four, human agency. We have to enforce our final decision making power to ensure the AI actually aligns with human values. Right. And what's fascinating here is the terminology she uses. She calls it intervention competence. Or even more bluntly, kill switch competence. Kill switch competence. That is a heavy phrase. It is. It's the literal human ability to manually stop automated processes. It's not just a big red button on a desk. It's the technical and operational authority to sever the loop when an AI deviates from safety parameters. And the research points out that this is the most dangerously underdeveloped part of the whole framework. Right. Massively underdeveloped companies are pouring billions into making AI faster and more autonomous. But almost nothing is going into the R and D to ensure humans can actually pull the plug. Which is why Professor Weiss makes this urgent plea. She argues that corporate R and D budgets must invest massively into ensuring humans retain this control. We have to build the brakes, not just the engine. Exactly. Because if we build systems that operate faster than our ability to monitor them, and we don't have the architectural authority to intervene, FOLO becomes a mathematical certainty. The core message here from Dr. Weiss is that the best answer to uncertainty is proactivity. Yes, there are so many possible futures of work, but ignoring the shift that just guarantees obsolescence. It really does. And I want to highlight a powerful closing sentiment from her article. She writes that progress is not just what new things we create, but also what we are spared from. The chances of success and failure have rarely been so close. It's up to us to build the guardrails. That is incredibly profound. And it brings me to my final thought for you, the listener. We just talked about strategic skill stacking. Right. Requiring three completely unrelated human skills to make yourself essentially irreplaceable. The human password. Exactly, the human password. So if you had to define your unique three skill password right now, that unique combination of intuition, social grace and chaos wrangling that no AI could ever replicate, what would it be? If you enjoy the podcast, please subscribe, leave a review and recommend it to your friends, family and colleagues.