AI Job Market for New Graduates: Career Ladder Problem

This is the AI Desk, where today's signals reveal tomorrow's power. And today's signal is terrifying. Every graduate with a LinkedIn account. That's almost all of them. Exactly, which is why we need to talk about it, because there's this question floating around right now in dorm rooms, family group chats, career offices, graduation parties, awkward brunches with uncles who still think, "Just walk in and hand them your resume" is advice. Classic uncle strategy. Should graduates be worried about AI? Like, actually worried. Not, "AI will change the future someday," worried. I mean, "I just spent four years getting a degree, and now the entry-level job I was supposed to get might not exist," worried. That is the real question. And I came into this episode ready to be optimistic. You know me. I wanted to say, "No, graduates are fine. Learn AI, be adaptable. Everything will work out." And then, you read the articles. And then, I read the articles. That usually ruins optimism. It really does, because I was reading the Wall Street Journal piece titled, AI is Wrecking an Already Fragile Job Market for College Graduates, and the whole premise is basically, companies used to give entry-level workers grunt work, and that grunt work was actually training. Now, AI can do a lot of that grunt work faster, cheaper, and without asking for dental insurance. That is the issue nobody wants to say out loud. The boring work was the training. Exactly. Mrs. Robinson's ghost is humming low. By the pool with a pink flamingo. They said, kid, don't stress. This episode is brought to you by MADCHITA and their new album WTF, Where Us The Forest? It's eco-pop engineered for the future, bold beats, global rhythms, and a message that actually matters. If you want music that hits your brain and your heart, explore WTF by MADCHITA. That's M-A-D-C-H-I-T-A. Streaming now on all major platforms. Get into plastics. It's fantastic. Floating in a sea of static. Nobody becomes a great analyst, designer, lawyer, marketer, or producer by immediately doing the glamorous work. You start with the messy work, the notes, the research, the first drafts, the spreadsheet cleanup, the junior edits, the, "Can you summarize this meeting?" tasks, the, "Pull 10 examples and make a deck by Friday" tasks. And those are exactly the tasks AI is getting good at. Right. So now the question is, if AI does the beginner work, how does a beginner become experienced? That is the broken ladder problem. Yes. The first rung is disappearing. And what makes it worse is that this is happening while the graduate job market is already rough. The New York Fed's labor market tracker showed that recent college graduates facing elevated unemployment and underemployment with recent grad unemployment around 5.7% and underemployment around 41.5% at the start of 2026. So, it's not just AI? No, that's important. It's AI plus hiring freezes, AI plus economic uncertainty, AI plus companies deciding they can grow without adding headcount, AI plus every entry-level job somehow requiring three years of experience, a portfolio, four internships, and the emotional resilience of a hostage negotiator. That is a very specific job posting. It is every job posting. And then I read the Bloomberg story, "New Grads Join Worst Entry-Level Job Market in Years," which says the class of 2025 faced closed doors as companies froze hiring, and AI made some lower-skill roles obsolete. So, graduates are entering a labor market where companies are hesitant to hire, and the roles they would normally hire for are being questioned. Exactly. And then, just to make everyone feel worse, Anthropic CEO Dario Amodei told Axios that AI could wipe out half of entry-level white-collar jobs, and push unemployment dramatically higher in the next few years. That quote hit hard because it came from inside the AI industry. Yes. It's one thing when random people online say, "AI is coming for jobs." It's another thing when the CEO of one of the most important AI companies says, "Hey, maybe we should stop sugarcoating this." And yet, I don't think the answer is panic. (laughs) Of course, you don't. I think the answer is precision. Graduates should be worried about the old path. They should not be hopeless about the new one. Okay, that's actually good. Graduates should be worried about the old path. Because the old path was simple, get credentialed, apply broadly, get an entry-level role, learn on the job, move up. The corporate Pokemon evolution. Exactly. But AI attacks that model at the bottom. It does not replace the senior person first. It replaces the tasks that justified hiring the junior person. And that's where this got personal for me, because my cousin graduated last year with a communications degree.Smart, hardworking, good writer, did everything right. She applied to marketing assistant jobs, social media coordinator jobs, content jobs, all the normal first-step jobs. And over and over again, the listings were insane. They wanted video editing, copywriting, analytics, paid ads, SEO, email marketing, Canva, Adobe, HubSpot, AI tools, three years of experience, and somehow, the salary was like, "You may also need a roommate and spiritual support." That is the compression. One entry-level worker is now expected to operate like a small agency. Yes. And what really got me was she said, "I don't even know what entry level means anymore." That sentence stuck with me. Because entry level used to mean potential. Now, it often means cheap generalist. Exactly. And AI makes that worse because employers look at a new graduate and think, "Can this person plus AI do the work of three people?" Which can be an opportunity if the graduate knows how to use AI well. But it's also unfair because how are you supposed to be experienced at using AI professionally if nobody gives you the first professional environment to learn in? That is the paradox. And I saw it with my nephew. He studied computer science. Two years ago, everyone told him coding was the safe path. The golden ticket. Right. Now, he's applying for junior developer roles and competing with thousands of applicants, laid-off tech workers, offshore teams, and AI coding assistants that can generate the boilerplate work he used to be hired to do. He told his mom, "I feel like I learned the entry-level version of a job that companies don't want anymore." That is brutal. It is. And I don't know how you tell graduates not to be scared when real people are feeling that. You don't tell them not to be scared. You tell them what to do with the fear. (laughs) Okay, professor survival guide. What should they do then? First, they need to stop thinking AI skills means typing prompts. Thank you, because every career advice post right now says, "Learn AI," and then refuses to explain what that means. Exactly. For graduates, learn AI should mean three things. One, use AI to produce real work, not demos, not cute prompts, actual outputs, a research memo, a market analysis, a prototype, a campaign plan, a financial model, a short video, a portfolio site, a customer support workflow. Receipts. Yes, proof. Two, learn how to verify AI because employers do not need graduates who blindly trust models. They need people who can catch hallucinations, check sources, test assumptions, and improve the output. That might be the most underrated skill. The AI gives you something that sounds confident, and you need the judgment to say, "This is nonsense in a blazer." Exactly. Three, learn the domain. AI makes shallow generalists cheaper, but it makes strong domain thinkers more powerful. Explain that. If you know nothing, AI gives you average output, and you may not know whether it is good. But if you understand a field, healthcare, law, education, construction, finance, design, local business, manufacturing, AI becomes leverage. The graduate who understands the work and uses AI to move faster has an advantage. The graduate who only knows how to prompt does not. So, the new resume is not, "I know ChatGPT." It's, "Here's what I built with AI, here's how I checked it, and here is the problem it solved." Exactly. That's actually useful, Rowan. And it connects to another article I read from the World Economic Forum. Is AI closing the door on entry-level job opportunities? The piece frames AI as reshaping the career ladder by putting some entry-level roles at risk while also widening access to global talent and changing what employers need. So, graduates are not just competing with AI, they're competing with everyone who uses AI well. Yes. That is the uncomfortable part. AI globalizes competence. That sounds like a villain line. It is just economics. Same thing, sometimes. Fair. But this is where I think the optimistic case exists. Graduates are often closer to the new tools than senior workers. They are less attached to the old workflow. They can adapt quickly. They can build portfolios faster. They can automate parts of their own job search. They can learn by doing. But only if they stop waiting for permission. Exactly. That's the part I keep telling my cousin. I told her, "Do not wait for someone to hire you before you start acting like the person they should hire." Make sample campaigns, audit a local business's social media, build a content calendar, make a fake launch strategy for a product you love. Use AI, yes, but then add your taste, add your judgment, add your understanding of people because employers are drowning in bland AI writing right now. If you sound human and strategic, that's an advantage. That is a key point. AI raises the floor, but it also makes the average more crowded. Yes, everything sounds polished now. Polished is cheap. Specific is valuable. That is the line. Write that down. Uh, we are recording this. Still, it deserves a dramatic pause. Polished is cheap. Specific is valuable.And that is true for graduates. The worst thing a graduate can do right now is become generic. Generic resume, generic cover letter, generic portfolio, generic, "I'm passionate about innovation." Straight to jail. The best thing they can do is show evidence of initiative. Real projects. Real projects. Real taste. Real taste. Namibia. Land of the cheetahs. This episode is brought to you by Mad Cheetah and their new album, WTF, Where is the Forest? It's eco-pop engineered for the future. Bold beats, global rhythms, and a message that actually matters. If you want music that hits your brain and your heart, explore WTF by Mad Cheetah. That's M-A-D C-H-I-T-A. Streaming now on all major platforms. Ahhh. Real curiosity. And real judgment. Okay, but let's debate this, because I know some people listening are thinking, "That sounds nice, but not everyone has time to build a portfolio for free. Some people actually need a job. Some people are working part-time. Some people are caring for family. Some people are burned out." That is true, and we should not turn structural problems into personal branding advice. Thank you. Because sometimes, the internet acts like every crisis can be solved by posting more on LinkedIn. It cannot. There is a real policy and employer responsibility here. Companies cannot eliminate junior roles and then complain about talent shortages later. Exactly. Who becomes senior if nobody gets to be junior? That is the long-term risk. If companies over-automate entry-level work, they may save money now, but destroy their future talent pipeline. Corporate America loves eating the seed corn. That is one way to put it. It's accurate. Companies need apprenticeship models for the AI era, not just internships where students fetch coffee or make decks nobody reads. They need structured roles where graduates learn how to supervise AI outputs, handle exceptions, communicate with clients, understand context, and build judgment. So instead of junior workers doing grunt work, they become junior operators. Yes. AI operators, workflow coordinators, verification analysts, domain researchers, client translators. The entry-level job may not disappear completely, but it may change shape. That is the hopeful version. And we are already seeing pieces of it. Some companies are not saying, "We don't need young workers." They are saying, "We need young workers who can use technology to solve problems." An article about Otis CEO Judy Marks made that point directly. Her message to grads was not to panic, because new graduates may have an advantage in knowing how to solve problems with technology. That's a good counterweight, because not every CEO is saying, "Replace them all with robots." Correct, but the burden is shifting. Graduates have to prove they can create leverage quickly. Which is scary, but also maybe exciting? Both? I keep thinking about my niece. She graduated with a psychology degree and was panicking because she didn't know how that translated into a job, and then she started using AI to analyze customer reviews for small wellness brands. She'd take public reviews, identify patterns, write a short insight report, and send it to local businesses. Not spammy, actually thoughtful. One owner responded and asked her to help rewrite their client intake process. That turned into paid work. That is a perfect example. She combined human understanding with AI-enabled analysis. Exactly. The AI helped with volume, but her psychology background helped her understand what people were really saying. That's the difference. That is what graduates should look for. Not, "What job title matches my major?" But, "What human problem do I understand better because of what I studied, and how can AI help me solve it faster?" That's much better than, "Follow your passion." Much better. Follow the problem. Yes, follow the problem. Okay, but what about graduates who are not entrepreneurial? Because not everyone wants to cold email local businesses and become a mini-consultant at 22. Some people just want a normal job. Then they need to interview differently. How? They need to show employers how they think with AI. For example, instead of saying, "I used ChatGPT for research," they should say, "I used AI to generate a first pass market map. Then I verified the top sources manually, found three errors, corrected them, and turned it into a client-ready summary." That sounds so much better. Because it shows process. Employers are nervous about AI slop. A graduate who can show disciplined AI use looks safer. So the new interview flex is not, "AI made this." It's, "Here's where AI helped, here's where it was wrong, and here's what I added." Exactly. That's actually a great framework. AI helped, AI was wrong, I added. (laughs) That should be on a mug. Maybe not a mug. A tasteful mug. Fine. Now, I want to talk about parents, because parents are panicking too. Yes. I had a friend ask me, "Should my kid even go into marketing?"And I didn't know how to answer, because the honest answer is, yes, but not the old version. That applies to many fields; marketing, coding, law, finance, journalism, design. The field may survive, but the beginner tasks inside the field may be transformed. So don't ask, "Is this major safe?" Ask, "What parts of this field are becoming automated and what parts still require judgment?" Exactly. For marketing, maybe AI writes first drafts, but humans still understand brand, timing, culture, emotion, trust. For coding, AI writes boilerplate, but humans still define architecture, debug complex systems, understand users, manage trade-offs. For law, AI summarizes documents, but humans still argue, advise, negotiate, and take responsibility. For finance, AI can model scenarios, but humans still interpret risk, incentives, and client behavior. So the safest graduates are not the ones who avoid AI. They're the ones who move up the judgment chain. Yes, move toward judgment, move toward context, move toward accountability- And away from being the person whose entire value is, "I can make a first draft." Because AI can make infinite first drafts. Infinite mediocre first drafts. Exactly. Okay, let's make this practical. If a graduate is listening right now and spiraling, what should they do this week? This week? Yes, not five-year strategy, this week. First, pick one field they are targeting, not 20, one. Painful, but necessary. Second, collect five job postings in that field. Highlight the repeated tasks and tools. Third, identify which tasks AI can help with. Fourth, build one small project that proves they can do those tasks better with AI. Example? If they want marketing, build a campaign audit for a real brand. If they want finance, create a simple market brief with sources and assumptions. If they want operations, design a workflow that saves time in a real process. If they want HR, create an onboarding plan and explain how AI could personalize it. If they want software, build a small product, document how AI helped, and explain what they personally designed and fixed. And then? Put it somewhere visible; portfolio, personal site, LinkedIn post, PDF, GitHub, whatever fits the field. Then use it in outreach. Not, "Please hire me." Instead, "I noticed this problem. I made this short example. I'd love to learn how your team thinks about it." That is so much better than, "I'm excited to connect." Everything is better than that. True. The goal is to become concrete. Concrete beats credential only. That's the new game. Yes, credentials still matter, but proof matters more than it used to. And what should parents stop saying? "Just get any office job." Yes. Also, "AI is just a fad." Definitely stop saying that. And, "They'll always need people," because, yes, they will, but they may need fewer people or different people or people with different skills. Better advice would be learn the tools, build judgment, make proof, and stay close to real problems. That was annoyingly clean. Thank you. I'm still worried though, because I don't want to pretend this is fair. Graduates did what society told them to do. They went to school. They took on debt. They built resumes. They applied, and now the rules changed mid-game. That is true, and that is why this is not just an individual adaptation story. It is a social story. Universities need to change. Employers need to change. Governments need to pay attention. AI companies need to stop pretending productivity gains have no labor consequences. And graduates need better advice than, "Network harder." Yes. They need a new apprenticeship model. Because the scariest version of this future is not that graduates never work. It's that they work but never learn. They become button-pushers, supervising tools they don't understand with no path to mastery. That is the danger; a labor market full of operators but fewer experts. And then ten years from now, companies say, "Why can't we find senior talent?" Because you automated all the junior talent before they could grow up. Exactly. So should graduates be worried? Yes, but they should be specifically worried, not vague doom, specific preparation. Worried enough to adapt, not worried enough to freeze. Exactly. That's the takeaway. AI is changing the first job, maybe even breaking parts of the first job, but it's also creating a new kind of first worker, someone who can use AI, question AI, improve AI, and bring human judgment to the output. And the graduates who understand that early may have a real advantage. Advantage. But the system has to meet them halfway. Yes. Because if we want experts in the future, we still need beginners today. That is the line. That's the episode. This is the AI Desk. Where today's signals reveal tomorrow's power.