AI raises unemployment, and this June, it’s hitting fresh graduates first. Every year, a new wave of U.S. college grads takes flight into the workforce like hopeful fledglings kicked from the nest. But now, that nest is wired with sensors, padded with synthetic feathers, and guarded by a tireless chatbot named Kevin. As millions toss their caps in celebration, they’re landing in industries that see them less as promising newcomers and more as expensive redundancies, easily replaced by artificial intelligence.
The numbers already have that vaguely metallic smell. Recent unemployment figures for college graduates have crept up to 5.8 percent in the United States. That might not scream crisis on its own but consider where it’s hitting hardest: technical fields like finance and computer science, ironically, the very places that taught young workers how to build the bots now elbowing them out of the way.
Oxford Economics put it bluntly: entry-level roles are evaporating faster in sectors where AI has made the most progress. It’s an observable phenomenon, the intern has been replaced by a virtual assistant that doesn’t sleep, complain about lunch options, or need mentoring. All it asks is for your data and maybe your job.
5 Key Takeaways
1. AI is pushing out fresh graduates. As AI adoption accelerates, industries that once welcomed junior talent now see new graduates as expensive and unnecessary. Entry-level roles, especially in fields like computer science and finance, are disappearing, replaced by tireless AI tools that cost less and require no training.
2. Companies are skipping human hires for automation. Firms are adopting “AI-first” hiring strategies that effectively mean “humans-last.” Junior engineers, analysts, and assistants are being replaced by models that can work around the clock. The message is clear: if AI can do the job passably, there’s no need to hire a person at all.
3. The AI hype hides serious risks. Builder.ai’s collapse is a cautionary tale. The company marketed itself as “AI-powered” while relying heavily on human developers behind the scenes. When the illusion crumbled, so did investor trust. This exposes the growing problem of “AI washing”, overselling automation to attract funding.
4. Entry-level roles are critical and disappearing. Thru elimination of beginner positions, companies not only reduce headcount but also dismantle the path to future leadership. Entry-level jobs are where people learn, grow, and gain experience. Without them, there’s no sustainable pipeline of skilled professionals.
5. Grads are adapting, but the system is broken. Many from the Class of 2025 are freelancing, building startups, or personal brands bypassing the traditional career ladder entirely. But this isn’t a sign of innovation; it’s a survival tactic in a job market that increasingly shuts the door before the first knock.
The New Corporate Motto: ‘Have You Tried AI-ing It First?’
What’s more chilling than hearing executives say they’re adopting an “AI-first” approach to hiring? Realizing that “AI-first” often translates into “humans-last.” One tech company declared it would no longer hire engineers below a certain rank, why bother, when LLMs can do junior-level coding? Another startup that once employed 75 data analysts now gets by with one.
Companies aren’t even pretending to hide it anymore. Automation is the name of the game, and AI is the player-coach-referee hybrid nobody asked for.
What once sounded like fevered Davos fantasy, “soon, machines will handle the grunt work!”, has become Tuesday at the office. Marketing analysts, finance associates, even research assistants are being told they’ve been pre-optimized right out of relevance.
For some firms, the lure is obvious: why spend six figures on a team of humans when one AI model, like Claude Opus 4, can code for hours without snacks or Slack messages? What it lacks in personality, it makes up for in performance metrics.
AI Raises Unemployment: AI Should Be Your Assistant
At The RegTech, a company based in Dubai with both feet firmly in the real-world regulatory trenches, we view AI not as a new manager, but as a rather bright intern who still needs supervision.
There’s real danger in confusing capability with autonomy. Yes, generative models can draft code, summarize contracts, even hold passable conversations about tax reform. But to throw humans out of the loop entirely is like giving the intern keys to the boardroom and then blaming them when the company stock nosedives.
AI should assist, not replace. Think of it as a power tool, not the carpenter. Letting it work unsupervised not only risks shoddy output but also hollows out the very structure of professional development. Entry-level jobs are where future managers learn how things work. If those rungs are sawed off the ladder, what’s left? A drop.
The AI-as-boss fantasy might work for a while. But when it fails, and it will, it’s the companies that kept the humans around who will be left standing.
Builder.ai and the Cautionary Tale of ‘Pizza-Level Software’
Speaking of AI dreams turning into pileups, let’s talk about Builder.ai, the poster child for what happens when hype outruns hardware. Founded in 2016 with a promise as digestible as a margherita slice (“software as easy as ordering pizza”), Builder.ai rode the AI marketing wave straight into a $1.3 billion valuation. It scored investment from Microsoft, cozy chats with the Qatar Investment Authority, and a flashy pitch: even non-coders could build sophisticated apps with a little help from their cheerful digital assistant, Natasha.
Except Natasha was mostly just smoke and mirrors. Behind the scenes, a small army of developers in India did the heavy lifting, while AI took the credit. It wasn’t so much intelligent automation as it was cost-effective outsourcing dressed in algorithmic clothing.
By 2025, the ruse collapsed. The company declared bankruptcy. Investors realized the tech was more marketing prop than miracle worker. Lenders seized millions. Employees were laid off en masse. Customers, many of them small businesses, were left scrambling to recover.
What died wasn’t just a company, but the illusion that slapping “AI-powered” on your product can substitute for actual, functioning intelligence.
AI Raises Unemployment: AI Washing, The New Greenwashing
Builder.ai’s collapse is fueling an overdue conversation about “AI washing”, the corporate habit of rebranding anything remotely technical as artificial intelligence in order to draw capital and attention. It’s the Silicon Valley version of putting a salad on the menu and calling your burger joint healthy.
The fallout isn’t just embarrassment for founders or red ink for investors. It’s operational chaos for companies that relied on hollow tech, and growing skepticism from a public starting to realize that not every AI startup is the next OpenAI. Some are just software shops with very loud megaphones.
And yet, despite the wreckage, the broader low-code/no-code market continues to expand. Gartner predicts that by 2028, 60 percent of new enterprise applications will be built on these platforms. The dream isn’t dead, it’s just developing a stronger nose for snake oil.
What’s a Graduate to Do When the Ladder’s Disappearing?
So where does this leave the freshly minted Class of 2025? Staring into an abyss lined with résumé-optimized bots and hiring freezes, some have chosen to skip the corporate ladder entirely.
Some are founding startups. Others are freelancing, coding on the side, or building personal brands before they can legally rent a car. There’s an urgency in their ambition, an awareness that the old playbook no longer works when the rules are being rewritten in Python.
But let’s be clear: this isn’t a healthy evolution. If companies increasingly treat junior roles as disposable, they’re not just alienating a generation, they’re stripping their organizations of future talent. There’s no middle management without a bottom to start from. There’s no “ten years of experience” without year one.
The Mirage of AI Autonomy
One recurring theme in conversations with startup founders and corporate executives is a quiet confession: these AI tools are good, but not great. They hallucinate, misinterpret, and require babysitting, constant training, and no small amount of handholding.
In other words, they’re not ready to take over the world. Not yet. And maybe never in the way we’ve been sold. The real danger isn’t that AI will replace every worker, it’s that companies will pretend it can. That pretense will justify layoffs, gutting of training programs, and a disinvestment in the people who keep businesses running long after the demo day buzz has faded.
And then, when things break, as they did at Klarna, the Swedish fintech that swapped customer service agents for AI chatbots only to bring back the humans months later, the same executives will act surprised. The future came too fast, they’ll say. Who could have predicted it would be so… buggy?
AI Raises Unemployment: Final Word from a Not-Yet-Obsolete Reporter
We’re not Luddites at The RegTech. We build regulatory technology, after all. But we also believe that good regulation starts with honesty, about what tech can do, and what it can’t. The truth is: AI raises unemployment. It does. Not across the board. Not forever. But enough to matter. Enough to worry. Enough to ask hard questions about what kind of economy we’re building and for whom.
We have a choice. Use AI as a tool that lifts workers up, or wield it like a wrecking ball to knock them out of the way. So far, too many companies are reaching for the latter. And if they’re not careful, they may find that when the bots fail, there’s no one left who remembers how to clean up the mess.
We are here to help governments, financial institutions, and businesses to effectively comply with growing regulatory requirements through technology.
