AI時代におけるリーダーシップの盲点

In 1951, philosopher Martin Heidegger told a small audience, “The most thought-provoking thing in our thought-provoking time is that we are still not thinking.” Few understood him then. Seventy-five years later, the observation has become unavoidable because AI has forced every leader to confront a question about the nature of intelligence and thinking itself. If thinking is nothing but what machines can do, only faster, we have no case: We outsource it to machines. But if thinking is something else — an embodied, attentive activity through which reality reveals itself — then leadership in the age of AI is the task of cultivating a generative capacity no machine can replicate.

Consider the doctor who treats screens instead of patients, the teacher constrained by standardized testing, or the World Cup referee whose real-time decisions are repeatedly overturned by a video assistant referee. Everywhere, situation-sensitive judgment is being replaced by what Hartmut Rosa calls execution logic: prestructured parameters that turn decision makers into mere executors.1 As spheres of discretion disappear, the creativity of human agency drains away. Beneath these surface symptoms sits the deeper question now beginning to surface in boardrooms: What is irreplaceable about us, and which intelligence will be the foundation of durable advantage once everything codifiable has been automated?

Every leader I work with — in business, government, international institutions, and nongovernmental organizations — reports the same thing. The machine is spinning faster than they can process and think. The acceleration extends far beyond AI: the inbox, the KPIs, escalating disruptions, tools meant to save time that consume more. Overwhelm has become a shared planetary experience. It is also an early warning signal that something essential is being eroded, precisely when we most need it.

This erosion has a name, a diagnosis, and a response. The name is intelligence monoculture: the assumption that AI is the only intelligence worth investing in. The diagnosis is that monocultures, sooner or later, collapse. Our response should be to create a second infrastructure, running in parallel to the agentic AI-enabled IT stack: a deep-sensing leadership infrastructure that cultivates the collective capacities to co-sense and cocreate at the level of the whole system. With it, AI becomes survivable and useful. Without it, the first infrastructure depletes the very soil it is rooted in — heading toward erosion and, eventually, collapse.

This is the blind spot. Leaders have a strong grasp of 何 they do (the actions they take, the strategies they execute) and どのように they do it (the processes, the systems, the tools). What remains hidden is the inner place from which they operate: the source of attention, intention, and creativity that no machine can replicate.

The age of AI forces us to clarify our assumptions about intelligence. Can thinking be reduced to computation and pattern recognition? Or is human thinking qualitatively different? And underlying this looms a deeper question: Who are we as human beings? Are we mere extensions of increasingly powerful algorithms — or genuine sources of awareness, intention, and agency?

Three Intelligences for the Age of AI

Intelligence is not one thing. At a minimum, three forms must be differentiated and integrated.

AI in the form of large language models (LLMs) is a pattern-prediction machine, matching and meshing existing human knowledge at a superhuman speed. Trained on existing data, AI deals extraordinarily well with dynamic complexity. It’s powerful — but structurally backward-looking, even when it appears to look forward.

Organic intelligence (OI) is the intelligence of structurally coupled living systems in ecologies of relationships. It senses multiple perspectives and orients to see with rather than just look at.2 This is where empathic listening lives. OI handles social complexity — the texture of multiple worldviews, cultures, and interests.

Source intelligence (SI) is the intelligence of the whole social field — the social soil from which all perspectives emerge. It is sourced not only from what is but from what is emerging. SI also stands for soil intelligence: the intelligence of the social mycelium running through that soil that connects what looks separate aboveground. Examples are entrepreneurs and leaders who sense and create a future that does not yet exist.

SI is grounded in what Eva Pomeroy and I have called fourth-person knowing: the source from which collective action arises.3 It handles emerging complexity: Where the solution is unknown, the problem keeps changing, and it is unclear who needs to be at the table.

The three intelligences are highly interwoven and nested, with source intelligence at the core and organic and artificial intelligences in the surrounding spheres. An intelligence monoculture — almost entirely dominated by AI — would look like an empty shell. There would still be some hardware. But the living, breathing inner core would be gone, turning the shell into a superhardened iron cage for those trapped inside.

The standard fear about AI runs one way: Machines are becoming more like humans. The real danger, though, may run in the opposite direction. We are becoming more like machines — not physically, but epistemically: We see thinking as computation, learning as data processing, creativity as recombination, decision-making as optimization, and the human self as algorithm. That epistemic conversion is what makes LLMs so seductive: They do not need to actually understand. They only need us to have already redefined understanding as what they do.

The Cave and the Sun

At the heart of the leadership challenge in the age of AI lies the question of where human attention, creativity, and agency originate. The late Bill O’Brien, a former CEO of Hanover Insurance, named it in a single sentence: The success of an intervention depends on the interior condition of the intervener.

In our work with teams across sectors, we have identified four structures of attention that organize how we listen, think, and act:4

1.0: Downloading. I listen to what I already know. Attention originates from inside the system; the interior condition is enclosed and reactive (ego-centric).

2.0: Factual listening. I lean into new facts with curiosity. Attention originates from the boundary of the system; the interior condition is transactional (object-centric).

3.0: Empathic listening. I see the world through the perspective of another. Attention originates from the field of relationships (relation-centric).

4.0: Generative listening. I listen to what is emerging from the edges, leaning into its best future potential. Attention originates from the surrounding sphere of potential; the interior condition becomes permeable to what wants to emerge (eco- or cosmo-centric).

The blind spot operates differently at each level. The arc from 1.0 to 4.0 is a shift in the structure of attention. What Plato names allegorically, leadership in the age of AI must name operationally. Prisoners chained in a cave see only shadows cast by a fire behind them; at levels 1.0 and 2.0, they take the shadows for reality. Much of today’s management lives among shadows — AI-generated projections, KPIs, dashboards, pattern matches mistaken for understanding. At Level 3.0, we turn around: We see the fire that casts the shadows. At this level, systems begin to see themselves. At 4.0, we step outside the cave into sunlight, into the realm of the source, which illuminates all things but cannot be seen by looking directly at it.

AI produces ever-more-convincing shadows. It simulates all four levels with astonishing mastery — text patterns that sound ego-centric, object-centric, empathic, even field-aware. But the simulation comes from patterns without interior condition — no witnessing awareness, no deep thinking. No one is there.

Perhaps the ultimate gift of AI is this: It holds up a mirror that forces us to see ourselves and ask, “Who are we? And who do we want to become?”

Four Levels of Collective Action and Strategic Innovation

Resilient organizations operate and innovate across four levels of collective action. Each level involves a distinct structure of attention and, in the age of AI, a distinct set of core leadership skills for the respective human-AI interface.

Level 1.0: Pattern-Executing — Automating. The first level is pattern executing and replicating: This operates with the logic of downloading, as in the levels of attention above. Agentic AI is an unprecedented driver of this level. The human-AI mode is delegation: AI or machine intelligence takes over well-defined cognitive tasks. Think about a fully automated production line. The core leadership skills here center around judgment, or how to recognize plausible but false AI outputs and results. The focus on automation can liberate human attention for higher-level work. This is where most investment flows today — and where Rosa’s execution logic operates in its purest form.

Level 2.0: Pattern-Adapting — Augmenting. The second level is pattern adapting and adjusting to the context of the environment. Human attention engages in object-centric ways — noticing disconfirming data, exceptions, and anomalies — but intention and agency remain within existing frames. The human-AI mode is navigating — Kasparov’s centaur: human strategist on top, AI as the powerful body underneath, with the human steering. This is what Nobel laureates Daron Acemoglu and Simon Johnson call “machine usefulness”: AI complements rather than replaces human labor.5 An MIT Media Lab study on cognitive debt found that LLM-assisted writers’ showed neural connectivity up to 55% lower than that of those who wrote without AI — and that sequence matters: Those who worked brain-first and then engaged AI showed significantly stronger metacognitive engagement than those who started with AI.6 At Level 2, the core leadership skills involve intention setting, sensemaking, and good judgment.

Level 3.0: Pattern-Shaping — Co-Sensing. The third level is pattern sensing and pattern shaping. Conversations shift from debate to reflective dialogue, that is, to thinking together. The move from sensing to shaping defines this level. Here, all three intelligences interact. OI tunes into the multiple perspectives at play. SI leans into emergence. AI surfaces patterns across large-scale data that no individual could perceive — and, used well, holds up a mirror in reflective dialogue that helps humans become more aware of their own assumptions and agency. The human-AI mode is shaped by a partnership with machines, revolving around orchestration and mirroring. This mode requires holding spaces for multiple intelligences to interplay, which in turn requires the core leadership skills of holding space for co-sensing, discernment, intention setting, and co-shaping to happen.

Level 4.0: Pattern-Originating — Deep Sensing and Cocreating. The fourth level is pattern-originating: deep sensing and cocreating. Here, SI moves to the core. Sensing what is shifts to sensing what emerges — the highest future potential. Reflective conversation shifts into generative dialogue: collective creativity and flow. The human-AI mode is holding the space: Origination emerges from human attention that becomes permeable to the field (eco- or cosmo-centric). AI moves from the center to the periphery, if it appears at all (a transcript, a reflective surface to return to later), and is not part of the originating act. The core leadership skills at this level center around holding space for deep sensing, moral discernment,7 shared intention, and cocreating.

In other words, the core leadership skills of the lower levels are included and recontextualized in the higher levels of collective action. One of the most critical leadership capacities today is the metacapacity to balance all four of these levels appropriately. Without that rebalancing, the gravity of AI pulls everything toward Level 1.0 and 2.0 monocultures.

From Machine to Living System

Industrial-era companies were designed like machines: standardized, process-driven, hierarchical, and replaceable. AI-era organizations, as my colleague Lili Xu has observed, increasingly resemble living ecosystems: dynamically collaborative, decentralized, adaptive, and responsive in real time. The most powerful companies of the future may not be the largest but the ones that learn and sense into emerging opportunities the fastest.

As AI dramatically reduces the cost of replicating expertise, what was once the source of competitive advantage — proprietary methods, scale, 10 years of training — collapses. What is truly irreplaceable about a company in the age of AI? Not algorithms; those are commodifying. The real source is the capacity to build organizations where technological intelligence and human field intelligence can evolve together.

The hidden infrastructure for this resilience, says Xu, is the people who sense tensions before they become crises, who hold trust across stakeholder groups, who perceive what customers cannot articulate. These forms of intelligence rarely appear in KPIs, yet they are often the source of an organization’s deepest competitive durability. This constitutes the paradox of the AI era: The more that intelligence becomes abundant, the more the relational and field-based intelligence becomes scarce — and therefore valuable.

What organizations now need to do is invest in deep-sensing infrastructure with the same seriousness they invest in AI. This is the other half of the infrastructure that is missing today in most organizations and economies.

For leadership teams ready to assess where they currently stand, the first question to ask is “How much leadership attention is currently going to the first and second levels of action, and how much to the third and fourth?”

Four mini diagnostics can help to clarify that picture:

• How much time in meetings is spent downloading and debating (levels 1.0 and 2.0) versus engaging in reflective and generative dialogue (levels 3.0 and 4.0)?
• Where is the center of gravity of how your organization currently operates: pattern-executing, pattern-adapting, pattern-shaping, or pattern-originating?
• To meet the demands of our age, which of those levels needs strengthening and more leadership attention now?
• What support structures — tools, practices, places — have you created that help your teams and organization to develop deep sensing and innovation capacities around levels 3-4?

Beyond the Blind Spot

Inside the cave, we take shadows for reality. AI-generated projections are mistaken for understanding. What is missing is the Level 3.0 capacity to turn around and see the structure that generates the projections — and the Level 4.0 capacity to step outside into sunlight, to originate new patterns from source.

AI produces ever-more-convincing shadows — depth, empathy, even wisdom — simulated from patterns without an interior condition: without the awareness that notices its own awareness. The current cave that we are operating in is our blind spot. Turning around and stepping outside requires what no AI can do for us: the cultivation of an interior condition from which we can see more deeply, more clearly, and more collectively.

Max Weber warned of modernity’s iron cage a century ago. Today, the cage has a new name: the 1.0-2.0 machine, supercharged by a trillion-dollar industry, the logic of inevitability, and the daily downloading that floods our calendars and shapes our attention. Each of us faces a choice: Get absorbed into the machine, or turn around and step outside. Choose what story of the future you want to be part of — and give AI the role it deserves: tool, partner, mirror or master. That move, if performed collectively, requires a new minimal enabling infrastructure: deep-sensing spaces that enable organizations to upgrade their operating systems and their capacities to levels 3.0 and 4.0.

Every day, leaders have two critical allocations to make: the allocation of attention, and the allocation of budget. What percentage of each is going into automation? Into navigation? Into orchestration? Into deep sensing and pattern origination? If your ratio is vastly out of whack, you already know what the next move should be.

The cave is comfortable. The shadows are mesmerizing. The logic of inevitability whispers that there is no alternative.

There is.

#Leaderships #Blind #Spot #Age