What AI genuinely cannot replicate and why it matters more than you think

This article is part of the Wisdom Gap series. You're reading part 2. Here is part 1

Starting with what AI can help build

Before naming what AI cannot replicate, I want to be clear about what it can.

Technical literacy, financial acumen, data fluency, research synthesis, clear written communication, project management, and analytical reasoning. These foundational professional capacities remain essential. Anyone entering the workforce today who dismisses them is making a serious mistake. AI is genuinely useful for developing and augmenting several of them, and the professionals who learn to work fluently alongside these tools will have real advantages over those who don't.

I am not arguing that young professionals should ignore these foundations. I am arguing that they are necessary but nowhere near sufficient AND that the sustainability transition specifically will rise or fall on five capacities that no model, however sophisticated, can build inside a human being.

Here is what I mean.

The first capacity: Trust as infrastructure

In all my years of operating inside companies, buying and selling businesses, and watching organizations succeed and fail, I have come to believe that trust is not a soft skill. It is infrastructure. It is the load-bearing structure underneath every meaningful collaboration, every systemic change effort, every coalition built across competing interests.

Trust cannot be generated quickly. It cannot be optimized. It accumulates through consistency over time. Through and by showing up when it costs something, keeping commitments when breaking them would be easier, and being honest in situations where honesty is inconvenient. It is built in the specific moments when someone could have taken the easier path and chose not to.

AI can simulate warmth. It can produce the language of relationships with impressive fluency. What it cannot do is accumulate a track record of costly choices made in favour of integrity over convenience. That track record is what trust actually is.

For the sustainability transition, this matters enormously. The changes required in energy systems, supply chains, financial structures, land use, and consumption patterns cannot be mandated from above. They require broad coalitions of actors who don't fully share interests but agree to move together. That kind of alignment runs on trust. And trust, in my experience, can only be built by humans willing to invest the time and absorb the cost of building it.

The second capacity: Judgment under genuine uncertainty

There is an important distinction between optimization and judgment that the current AI conversation tends to blur.

Optimization is what happens when the variables are known, the constraints are defined, and the goal is to find the best path through a structured problem. AI is extraordinarily good at this. Better than humans in most domains and getting better quickly.

Judgment is what happens when the map doesn't exist yet. When the variables are unknown, the constraints are shifting, and the decision has to be made anyway because waiting for certainty is itself a choice with consequences. Every significant sustainability challenge I have encountered lives in this territory. Not because the science is unclear, but because the path from the current reality to the necessary future runs through human systems that are messy, contested, and resistant in ways that no model fully captures.

The leaders who will navigate the sustainability transition are not the ones who can optimize within known parameters. They are the ones who can act wisely in conditions of genuine uncertainty. Reading incomplete signals, weighing incommensurable values, and accepting responsibility for outcomes they cannot fully predict.

That judgment is not innate. It develops through experience with real decisions that had real consequences and couldn't be undone. It requires, in other words, formation.

The third capacity: Holding complexity without collapsing it

There is a profound human tendency, under pressure, to resolve complexity into simplicity. To pick a side. To reduce a nuanced situation to a clean narrative that feels actionable even when the real situation resists clean action. This tendency is deeply human, and it is also, in my view, one of the most dangerous failure modes available to sustainability leaders.

The sustainability transition is not a simple problem with a complicated solution. It is a genuinely complex problem. One in which the components interact in non-linear ways, where interventions in one area create unexpected consequences in others, and where the right answer in one context may be actively wrong in another.

Holding that complexity without collapsing it into false certainty requires a specific kind of cognitive and emotional capacity. The ability to sit with contradiction long enough to find a real path through rather than a convenient one. The willingness to say "I don't know yet" in situations where others are demanding answers. The discipline to keep multiple competing frames active simultaneously rather than prematurely committing to one.

AI is, by design, oriented toward resolution. It produces answers. The capacity I am describing is almost the opposite. It’s the willingness to remain productively uncertain and to let the complexity breathe until genuine clarity emerges.  As opposed to choosing the manufactured clarity.

That capacity is forged in the messy middle of real experience. It cannot be prompted into existence.

The fourth capacity: Genuine relationship across difference

The sustainability transition will never be achieved by people who already agree with one another.

It requires building genuine working relationships, not performative ones, across sectors, cultures, worldviews, and interests that are in real conflict. Between the engineer and the activist. Between the finance executive and the community organizer. Between the Global North and the Global South. Between generations with fundamentally different relationships to the crisis and its costs.

This is, I have found, among the hardest human work there is. It requires the ability to genuinely understand a perspective you fundamentally disagree with. NOT to perform understanding, NOT to find the diplomatic language that bridges the gap. Rather than actually inhabiting enough of someone else's reality to work with them effectively across the differences.

AI can generate the language of empathy with remarkable fluency. It can produce culturally sensitive communication, identify points of potential agreement, and suggest bridging framings with impressive sophistication.

What it cannot do is be genuinely changed by an encounter with another human being. The relational capacity I am describing is not a communication technique. It is what happens when two people bring their full humanity into contact with each other over time and allow themselves to be affected by what they find. That is how trust gets built across differences. And it is irreducibly, categorically human.

The fifth capacity: Long-horizon thinking as a felt reality

I have saved the most important for last. And in my experience, it is also the rarest.

Understanding intellectually that the long-term matters is not the same thing as being genuinely motivated by outcomes you will not personally live to see. The first is a cognitive position. The second is a felt reality. It is a deep orientation that shapes decisions at the level of instinct, not just analysis.

Most of the systemic failures I witnessed in business came down to this gap. Smart people who understood perfectly well that their short-term decisions had long-term costs, and who made those decisions anyway because the long term was abstract and the short term was real and immediate and tied to their compensation.

The sustainability transition requires — at scale, across sectors, in boardrooms, government offices and investment committees — people for whom the long horizon is genuinely felt. For whom the well-being of people they will never meet, in decades they will not see, registers as a real constraint on present decisions.

That orientation does not come from analysis. It comes from a particular kind of formation. From having been shown, by someone whose life demonstrated it, that it is possible to care genuinely about what comes after you. From having worked inside organizations where that caring shaped real decisions. From having absorbed, through proximity and experience, a relationship with time that extends beyond a quarter or a career.

This is perhaps the most human thing on this list. And it is the thing the sustainability transition needs most.

The argument, plainly stated

AI will continue to absorb the technical, analytical, and cognitive tasks that have historically filled the early years of professional careers. That process is accelerating and not reversing.

What it cannot absorb is the formation that happened around those tasks. The trust was built through consistency over time. The judgment developed through real decisions with real consequences. The capacity to hold complexity, build genuine relationships across difference, and care about outcomes beyond your own horizon.

These are not soft skills at the margins of professional life. In my view, they are the core competencies the sustainability transition specifically requires, and they are precisely what the wisdom gap threatens to eliminate.

The question that follows, and it is an urgent one, is what we do about it. Where does formation now occur for a generation entering a world in which the traditional container has largely been removed?

That is the question Article 3 takes up directly.