Editorial: AI’s Terrible Teens

Child prodigies have always fascinated us. The thrill isn’t just in what they can do, but in what they might become. If Mozart was composing sonatas at four, imagine what he could do at forty. But as history often reminds us, wunderkinds don’t always live up to the hype. Some burn out, others plateau. And some hit puberty like a wrecking ball — unpredictable, temperamental, and exhausting for everyone involved. At that point, anything is possible: triumph or tragedy, genius or self-destruction. But one thing’s for sure — it’s never easy.

The world’s newest prodigy isn’t a boy wonder with a violin. It’s artificial intelligence. From its humble beginnings as a party trick for end users, AI has grown at breakneck speed into a fixture of the modern workplace. It’s now firmly embedded in office routines, academia, and even journalism. But behind the scenes, AI is no longer a charming child—it’s a rebellious teenager. Still brilliant. Still full of promise. But now the question is no longer what it can do— t’s what it’s becoming.

Killer of Bullshit Jobs

First up: the technical tantrums. AI hallucinations—confidently wrong outputs—are no longer rare embarrassments; they’re recurring nightmares. One culprit is what some researchers have called ‘AI inbreeding’—new models trained on data polluted by earlier ones. That creates a downward spiral where machines copy machines, losing accuracy with each generation. While safeguards have improved, the issue has never been fully resolved and demands growing oversight with each new release.

Adding to the mess: hardware bottlenecks. Traditional silicon chips are hitting their thermal and performance limits. Data centres are cooking themselves silly, forcing throttled performance and sky-high energy costs. Alternatives like superconductors or neuromorphic chips exist—but they’re expensive, exotic, and not yet fit for the real world. Quantum computing holds long-term promise, but as of now,  it’s more pitch than prototype—bold slides, little results.

Yet AI is already sharpening its knives for the so-called ‘bullshit jobs’—a term coined by anthropologist David Graeber for white-collar roles that even their holders admit are pointless. Academic precariat jobs and bureaucratic busywork are ripe for replacement. In a market obsessed with efficiency, AI’s march seems inevitable. But if the machines take over, where do all the degree-wielding paper-pushers go?

The short answer: the public sector. Surveys—including one from McKinsey and recent OECD data—show a growing preference among university graduates, especially post-COVID, for government jobs. Safe, stable, and immune to market swings. But here’s the uncomfortable truth: many wouldn’t survive in the private sector with their soft degrees. And the public sector, too, is far from immune.

Behind closed doors, civil servants in European municipalities quietly admit that AI could already handle about half their daily workload. In smaller, cash-strapped regions, automation could soon look less like a luxury and more like a necessity. When that happens, large chunks of local government jobs—especially in clerical and administrative fields, many held by women—could vanish.

That shift won’t just affect employment statistics. It may reshuffle gender roles and up-end the already wobbly balance of domestic labour. If many women find themselves pushed out of the workforce, they might retreat into traditional care-giving roles—not out of nostalgia, but necessity. That shift could, in turn, give birthrates a gentle push upward. Then again, the loss of career prospects and income could just as easily accelerate disillusionment and delay family formation further. As with any teenager in turmoil, predicting what comes next is more art than science.

The Future is Decentralized

As AI stumbles towards adulthood, the question looms: who’s raising it?

In the US, the answer is increasingly clear—a handful of tech firms with government support have near-total control. In China, the state runs the show, with a tight grip and a clear agenda. Europe, meanwhile, is dithering over one-off investments in data centres, hoping to at least appear like it’s in the race. Spoiler: it isn’t. What is becoming clear, however, is that governments and markets are starting to look eerily aligned in their ambitions for AI. Whether in Washington, Beijing or Brussels, the Zeitgeist favours top-down control and uniform oversight—raising thorny questions about democratic freedoms, decentralisation, and human dignity in the age of the algorithm.

But how does one keep pace without sacrificing these values and freedoms? Perhaps the most important shift won’t come from speed or scale. It might come from direction. Rather than endlessly chasing smarter, faster models, the next leap could lie in how AI is deployed. Will we rely on AI agents to do tasks for us? Or AI tools that support us? Could decentralised AI—light, efficient, and open-source—become the new normal, especially in places locked out of the billion-dollar computer arms race?

AI is growing up. That doesn’t mean it’s going full Skynet, nor that it will fulfill every utopian dream. It means taking stock of its strengths, weaknesses, and place in the world. The trick will be to steer clear of the two classic traps of Wunderkinder: believing the hype too early, or writing them off too soon.

With realistic expectations and smart guardrails, some of which could be based on the upcoming teachings of Pope Leo XIV on the matter and a related revamp of subsidiarity-based discourse, AI could become more than a miracle—it could become genuinely useful. But without that, it risks joining the long list of over-promising prodigies who collapse under the weight of their own myth.

Statement

AI is no longer the obedient Wunderkind; it’s the surly adolescent of the tech world—volatile, brilliant, and dangerously self-assured. Once a novelty, it now risks imploding under its own hype. Technical dysfunctions like hallucinations and overheated data centers signal deeper structural flaws—some of them self-inflicted. As AI starts gutting “bullshit jobs,” the fallout is social as much as economic, threatening public sector stability and gendered labor patterns. Power is consolidating fast: Silicon Valley scripts the future while Europe stammers. Unless we shift from scale to intent—toward decentralised, supportive AI—we risk raising a genius that can’t stop sabotaging itself.