The sustainability crisis nobody is naming

This article is part of the Wisdom Gap series. You're reading part 1.

The wrong existential threat

The sustainability conversation has, for decades, been focused on carbon. On biodiversity loss, regenerative agriculture, circular economies, and the urgent transition away from extractive business models. All of it matters. The science is sound. The urgency is real.

But here's what I keep turning over in my mind: none of it will matter if we hollow out the very thing that makes civilization worth sustaining. In other words, the human capacity for judgment, relationship, wisdom, and formation.

Artificial intelligence is the most powerful business productivity tool I have seen in forty years of operating inside companies. Deployed thoughtfully, it genuinely could accelerate the transition to sustainable systems. Deployed the way most businesses are currently deploying it — as a pure efficiency play, a headcount reduction strategy, a way to compress the cost of entry-level work to near zero — it is quietly producing something that no climate model is currently measuring.

A generation of unformed humans.

Over the next thirty years, if we continue treating AI as an optimization engine rather than a civilizational inflection point, we won't need to debate whether our business models are regenerative. We won't have the humans capable of imagining, building, or running them.

That, as best I can see it, is the sustainability crisis nobody is naming.

What entry-level work was actually doing

To understand what's being lost, I think we need to be honest about what entry-level work was never really about.

It was not, at its core, about the tasks. The reports filed, the data entered, the calls answered, the meetings scheduled. Those were tasks that were the visible surface of something far more important happening underneath. The first ten to fifteen years of a professional career have historically functioned as a developmental container. A structured environment in which young humans learned, mostly by proximity and consequence rather than by instruction, how to be in the world with other humans under pressure, over time.

Consider what that container was actually building.

Discipline through consequence. Showing up when you didn't feel like it. Meeting a deadline because someone was depending on you, not because an algorithm sent a reminder. Learning the difference between effort and result in conditions that had real stakes attached.

Social calibration. Reading a room. Navigating hierarchy, conflict, and collaboration simultaneously. Beginning to understand that what someone says and what they mean are often very different, and slowly developing the sensitivity to close that gap.

Accidental mentorship. Watching a seasoned leader handle a crisis with grace and filing it away. Watching a different leader handle the same kind of situation badly and filing that away, too. Absorbing, through sheer proximity, decades of pattern recognition that the experienced person couldn't have articulated if asked. Yet transferred anyway, the way culture always transfers, through observation and osmosis rather than instruction.

Identity formation under pressure. Discovering what you actually believed when it cost something to say so. Building the scar tissue from decisions that didn't go the way you planned, and learning to live with it in a context where people remember.

None of this appeared in a job description. None of it showed up in a performance review. And almost none of it is being considered in the current conversation about AI and employment, which remains almost entirely focused on tasks.

The loss is not about tasks. The loss is about formation.

The efficiency trap — again

The business logic driving current AI adoption is straightforward and, on its own terms, impeccable. Entry-level cognitive work — the research, synthesis, first-draft generation, data processing, and pattern matching across large datasets — is exactly what large language models do well, quickly, and at a fraction of the cost of a junior employee who needs managing, makes mistakes, requires benefits, and takes eighteen months to become genuinely useful.

From a quarterly earnings perspective, this is an obvious win.

From a civilizational perspective, it looks to me like the latest iteration of a mistake we keep making. Optimizing local efficiency at the cost of systemic resilience.

We have done this before. We optimized agricultural supply chains to reduce costs and eliminated the redundancy that kept communities fed when the primary system failed. We optimized financial systems for return and eliminated the friction that had imperfectly slowed contagion when one node collapsed. We optimized organizational structures for productivity and eliminated the middle-management layer, which, despite its inefficiencies, had also served as a knowledge-transfer mechanism between generations of workers.

In each case, the efficiency gains were real and measurable. The systemic costs were diffuse and slow-moving and therefore easy to discount until they weren't.

The AI-driven elimination of entry-level work follows the same pattern. The gains are immediate and legible. The costs, like a generation entering mid-career without having built the relational, judgmental, and adaptive capacities that the entry-level container quietly produced, will take fifteen to twenty years to fully manifest. By which time, the decision-makers who accelerated this will have long since moved on.

This is not a new story. It is the extractive logic story applied to human formation.

What I'm calling the wisdom gap

What we are creating, systematically and largely without intention, is a wisdom gap.

Not a skills gap. The reskilling conversation, well-intentioned as it is, still operates entirely within the efficiency frame. How do we make young people useful in an AI-augmented economy? That is a real question. But it is not the most important one.

The most important question, as I see it, is where does formation happen now?

Wisdom, the kind that matters for building sustainable systems, leading organizations through genuine uncertainty, and making decisions whose second and third-order consequences require human judgment rather than optimization, is NOT a skill. It cannot be trained in a course or acquired through a certification. It accumulates over time, through consequence, relationship, and the specific kind of learning that only happens when you are responsible for something real, the stakes are real, and the humans around you are complicated and unoptimized.

That is what the entry-level container provided. Not efficiently. Certainly not consistently. But reliably enough, over enough generations, that we confidently built the professional and civic infrastructure of modern civilization on top of it.

Remove the container without replacing it, and you do not just change the economy. You change the quality of human beings that the economy produces.

One layer deeper than the conversation currently goes

The economic disruption argument is real and well-documented. The acceleration is measurable. Leading AI safety researchers at METR have found that the complexity of tasks AI can complete is doubling roughly every six months, a trajectory that makes Moore's Law look cautious. In 2019, early AI models could barely string together a coherent paragraph. By 2023, large language models were passing the Bar Exam and the US Medical Licensing Examination, outperforming ninety percent of human candidates. The displacement is already underway, and those who dismiss it are not paying attention.

I am not disputing any of that. What I am arguing is that the conversation has stopped one layer too shallow. It stopped at the economic surface, where the questions are about jobs, reskilling, and individual survival strategies. The deeper civilizational cost is still largely unnamed. Because what is disappearing alongside the entry-level jobs is not just income. It is formation. And that loss does not show up in any labour market report, any productivity metric, or any AI adoption ROI calculation currently being run.

That is the layer this series is about.

Why this belongs in the sustainability conversation

The sustainability and regenerative business communities have spent decades arguing that short-term extraction is not only ethically problematic but also strategically incompetent. Because it destroys the very systems on which long-term value depends.

That argument is correct. And in my view, it applies with full force to the current deployment of AI in business.

We are extracting efficiency from a developmental system without accounting for what that system was producing. In the language of regenerative economics, we are treating formation capital as an externality.  Which means it is a cost to be eliminated rather than an asset to be stewarded.

The carbon we can measure. The biodiversity we can measure. The wisdom gap we are creating will be measurable, too. Eventually. In the quality of leadership, the resilience of institutions, and the capacity of human beings to build and sustain the regenerative systems we keep saying we want.

In my experience, the sustainability community has always understood that what you measure determines what you manage. We are not measuring this. We are barely even naming it.

The first effects will be visible within a decade. The full civilizational cost will take thirty years to arrive. Which is precisely the same window we have to complete the sustainability transition.

What comes next

I don't have all the answers here, and I want to be clear about that. What I have is forty years of watching organizations optimize themselves into brittleness, and a growing conviction that we are doing it again, at a civilizational scale, with the most powerful tool we have ever built.

The questions I think we need to be asking are: What does deliberate formation look like when the traditional container is gone? What can AI genuinely not replicate, and how do we make those capacities the explicit priority they should have been all along? What do those of us further along in our careers owe the generation coming up behind us?

The sustainability community is uniquely positioned to lead this conversation. You understand systemic thinking. You understand externalities. You understand the difference between what is measurable now and what matters over time.

This is that conversation. And it is long overdue.