“I do not underestimate the importance of the left hemisphere’s contribution to all that humankind has achieved, and to all that we are, in the everyday sense of the word; in fact it is because I value it, that I say that it has to find its proper place, so as to fulfil its critically important role. It is a wonderful servant, but a very poor master.”

Iain McGilchrist, The Master and His Emissary, p. 437

This is the article that tries to chart a way forward as we continue with AI.

A brief recap, for readers arriving here from article 1/5.

The series has argued that AI is the most powerful instance yet of a particular mode of thinking that has been quietly dominant in our culture for centuries. Iain McGilchrist calls it the left-hemisphere mode. It works by stripping context, by operating only on what can be articulated, by producing confident outputs without the answerability that would let those outputs be checked against the world. Three traits explain how AI is left-hemisphere thinking run wild: decontextualization, language-dependence, misplaced confidence. Articles 2/5, 3/5, and 4/5 developed each one, with examples from my professional exchanges with AI (that readers likely have encountered themselves). The cumulative argument is that AI cannot fix these limitations from inside its own mode. The limitations are inherent to what AI is.

These limitations are not a bad thing. In fact, they are the opportunity for humans to leave their mark.

The two failures of the current discourse on human collaboration with AI

The current discourse on human-and-AI collaboration is lacking in at least two ways.

The first is thinness. A common word for the human contribution is “judgment.” We are told that AI does the heavy lifting and humans bring “judgment.” But judgment is a thin word for what humans actually have to do. It is a placeholder, not a description. It doesn’t tell you what to cultivate, what to protect, what to watch for in yourself. It is especially difficult to instill in junior contributors who are most at risk of lazy thinking attributable to AI use.

The second failure is positioning. The dominant frame for the human role is “human as a check,” which sits the human at the end of the AI’s work, inspecting outputs for failure. This is the frame baked into “human oversight,” “human review,” and much of the guidance organizations are currently writing about AI use. It positions the human as quality control: reactive, late-stage, watching for errors after the fact.

Both framings make the human’s role passive and opaque. Judgment is what you apply at the end; oversight is what you do at the end; review is what you do at the end. The human is the last filter before the work goes out. The AI is the source; the human is the check.

The whole argument of this series points the other way. If decontextualization, language-dependence, and misplaced confidence affect everything AI does, the human contribution cannot be a final filter. By then the work has already been done using an incomplete perspective. The human has to be present from the beginning, holding the perspective the AI cannot hold.

ACTS as the human stance, in action

McGilchrist gives us a richer vocabulary, and the right-hemisphere faculties described in articles 2/5, 3/5, and 4/5 give us a usable framework. I’ve summed them here using the acronym of ACTS:

The human ACTS. The framework names a stance, an action, a way of attending.

Two paragraphs on what the ACTS stance represents, before I get into practice.

McGilchrist’s right hemisphere is, among other things, the seat of how we attend to the world. His central claim about attention is that it isn’t neutral. The kind of attention you bring to anything determines what shows up. Narrow, instrumental attention shows you a world of objects to be used. Open, vigilant attention shows you a world of situations, relationships, and contexts in which the objects are placed.

The human contribution to AI-assisted work is, in this sense, a way of attending that the AI cannot perform. The AI brings narrow, instrumental attention to the tokens in front of it. That is what it is for. The human brings the broad, situated attention to what those tokens mean in the world they came from, what is in this situation that wasn’t in the training distribution, what is not yet visible because it cannot be put into words. The stance is to keep that broader attention present throughout the work, holding the wider field that the AI cannot see. It is the master’s stance from McGilchrist’s fable. The emissary’s narrow attention has become the architecture of AI; the master’s stance is what stays with the human.

ACTS in practice: the individual

The most practical form ACTS takes is self-questions. Not prompts you give the AI, but questions you ask yourself while working with AI. Questions to the AI just relocate the work back to the AI, which is the problem we’re trying to solve. Self-questions keep the work on the human side, which is the point.

Questions such as:

These are not a checklist. They are habits of awareness as you work. The practice is recognizing what you bring, not what you can extract. If you answer the attention question and notice you have been outsourcing attention, you stop. You read the AI’s output more slowly. You ask what you actually came here to do. If you answer the context question and notice the situation has features the training distribution didn’t, you make those features explicit (to yourself, not to the AI) before you decide how much weight to give the AI’s output.

There is a corollary to this practice that becomes clearer once you have started doing it. ACTS applies not just to how you work alongside AI but to which tasks you ask it to do in the first place. The highest-value AI uses I have found are not final products but highly-falsifiable research processes built around controlled datasets and specific questions I can review for myself. The AI does the legwork; the practitioner holds the question, the context, and the stake. Where the work in front of you can be structured that way, ACTS holds naturally. Where it cannot, the framework is asking you whether AI should be doing the work at all.

The faculties take years to develop in any practitioner. They are vulnerable to AI in a specific way: the AI is fast, the AI is fluent, the AI is always available, and the path of least resistance is to let the AI’s framing become yours. The thing you watch for, in yourself, is the slow ceding of these faculties. You watch for the moment you stopped reading the source document carefully because the AI summarized it. You watch for the moment you stopped trusting your own felt sense of a piece of work because the AI’s confident gloss made yours seem like an objection. You watch for the moment you let the AI’s draft be the version you defended, without quite noticing that you no longer felt able to say what it said with your own voice.

What you don’t delegate is the part of the work that can only be done by someone with stake. That is most of what makes the work yours.

ACTS in practice: the organization

The same framework changes shape at the organizational level.

Most organizations are absorbing AI as a productivity play. Efficiency, automation, cost reduction, throughput. It is itself a left-hemisphere framing, and the same set of problems follows from it as follows from any left-hemisphere framing applied to a situation that requires right-hemisphere holding: the parts get optimized, the whole is disregarded, and the costs land somewhere downstream that no one is measuring.

A right-hemisphere-respecting integration of AI looks different.

First, there are roles whose job is to hold what AI cannot hold. These differ from AI oversight roles, which are quality control at the end. The roles I mean have attention broader than what AI can articulate: people who hold the client relationship, who hold the institutional memory that isn’t documented, who hold the felt sense of where the firm is heading. These roles need to be staffed by people with stake and protected from being recategorized as bottlenecks when the productivity numbers get tight.

Second, work has to be structured so the right-hemisphere faculties get exercised rather than eroded. The faculties were developed, in previous generations of practitioners, through apprenticeship. The junior person did the work themselves, made mistakes, got corrected by someone more senior, and built up a felt sense of the work over years. AI threatens that apprenticeship by giving juniors a tool that produces work products that look polished from day one. The juniors don’t develop the muscles. The seniors lose the apprenticing relationship. Five years from now, there is no one with the right-hemisphere faculties the firm needs, because no one had to develop them.

Third, training has to be redesigned to take this into account. The relevant training cultivates the faculties AI cannot supply: structured reading of original documents (not AI summaries), client interactions held by juniors, the kind of slow attention that builds the felt sense over time. This needs to be a deliberate investment because the path of least resistance points the other way.

Organizations face a choice they are not currently treating as a choice. The choice is whether to absorb AI as pure productivity gain (and quietly let the right-hemisphere faculties atrophy across the workforce) or to absorb it as a tool whose limitations require structural compensation. The first looks cheaper in year one and more expensive in year five. Most organizations are heading for the first by default.

ACTS in practice: the advisor

This is where the argument gets sharpest, because the advisory craft is largely right-hemisphere work, and AI is more disruptive to it than to most kinds of work.

What advisors are actually paid for, when you strip the surface activity away, is context-holding (the felt sense of this client, this situation, this moment), the tacit (the craft sense that develops over years of advising), and stake (the answerability that makes advice trustworthy because it can be challenged later). These are the ACTS faculties precisely. The advisory craft is ACTS made into a job description.

AI is good at the things that look like advice from outside: producing articulate analysis, organizing material, framing options, drafting recommendations. It is bad at the things that make advice valuable from inside: knowing what this particular client needs that another client wouldn’t, knowing when a recommendation that looks correct on paper would fail in this room, knowing what you are answerable for.

The result is that AI use in advisory work is not a neutral productivity question. It is a question about what the advisor is still bringing that makes the advice theirs. As article three argued, tacit knowledge is what makes an expert an expert; careless AI adoption erodes that knowledge, and with it the substance of expertise the client is paying for. If the advisor adopts AI without changing anything else, what they end up bringing is the prompt, the editing, and the relationship management. The substantive part of the advice has been produced by something that has no context, no tacit feel, and no stake. What ends up reaching the client is branded AI output. The client may not notice immediately, but they will notice eventually, and the trust the advisor’s name has accumulated over years gets quietly burned.

The senior advisor and the junior advisor need to use AI differently, for reasons the framework makes explicit. The senior has the faculties; the AI is a tool that extends them, and the senior’s attention, context-holding, tacit sense, and stake are all in play. The junior does not yet have the faculties; the AI does for them what they need to be doing themselves in order to develop the faculties. If juniors use AI the way seniors do, the path to becoming a senior is broken. They become permanent juniors with a powerful tool, and the firm’s bench is hollowed out.

A closing thought

To close the series, I want to come back to where I started.

Years of sustainability work taught me to notice when a single mode of thinking has stopped serving the problem it is supposed to solve. In corporate sustainability, the dominant left-hemisphere thinking reduces whole-systems questions to metrics and calculations and then wonders why the calculations don’t deliver the outcomes the system actually needs. The pattern is familiar. The pattern is the same one McGilchrist describes happening across centuries of cultural history. The pattern is the one I have been arguing, across this series, is now showing up most powerfully in AI.

The framework I have offered, ACTS, is small. It won’t fix everything. What it does is give people who already sense something is wrong with the current discourse a place to stand. There’s a better way to articulate “judgment”. Human as a check was the wrong position. The work is bigger than that, and it is yours to do, before the AI gets to it.

McGilchrist’s line of caution is that the left hemisphere is a wonderful servant and a very poor master. The same is true of the most powerful thing the left hemisphere has ever built. AI is a wonderful servant. It cannot be the master. The role of the human, in the era of AI, is to be a master worth serving.

“Although the left hemisphere does not see and cannot understand what the right hemisphere understands, it is expert at pretending that it does, at finding quite plausible, but bogus, explanations for the evidence that does not fit its version of events.”

Iain McGilchrist, The Master and His Emissary, p. 234

I want to start with an exchange that may sound familiar to you if you’ve worked with AI a bit. It came up while I was preparing materials for a presentation on the electronics industry’s environmental footprint, and an AI was helping me build the supporting visuals.

The chart

The AI produced a bar chart comparing four industries on what it labeled “E-Waste.”

I read it and for a moment was impressed that the AI produced a coherent bar chart. Then I caught myself and looked again, because I was genuinely confused. There is no way fashion produces more e-waste than electronics. E-waste is electronic waste: phones, computers, screens, the discarded hardware of the electronics industry. Fashion produces textile waste, water pollution, and carbon emissions. The figures shouldn’t be anywhere near the electronics industry.

I sent the AI a single message: “it doesn’t make sense that the fashion industry produces more e-waste than the electronics industry. Please check that source.”

That was the right-hemisphere check. The bits-and-pieces that I see in front of me do not match the world. Please look again.

The reframe

Here is what came back

The AI’s response began: “You’re absolutely right to question this. But I want to clarify what was actually stated in the presentation. The comparison to the fashion industry was about carbon emissions, not e-waste production.” It went on to explain that the comparison between electronics and fashion was about greenhouse gas emissions (both industries contribute approximately 4% of global GHG emissions, roughly equivalent footprints), and that this comparison was specifically about carbon emissions and carbon footprint, not e-waste generation.

Read that response carefully. The chart was in front of me, clearly labeled “E-Waste,” showing fashion higher than electronics on whatever the y-axis was tracking. The AI’s response was essentially: the comparison was not actually about e-waste; you misunderstood. The confusion, the AI suggested, arose from “the proximity of these two different discussions” (e-waste statistics and carbon footprint) within the same section.

This is the reframe. The AI produced an output that was wrong. When the output was challenged, the AI located the problem in the user’s interpretation rather than in the output itself. It re-framed what the chart was “actually” doing in a way that conflicted with what the chart said on its face. It apologized “if the wording created this misunderstanding.” The apology was for a wording problem, which is subject to user interpretation, not for an accuracy problem in the chart itself. The accuracy of the chart was never in question, according to the AI.

The hedge

So I pushed back more directly. My second message: “In the bar chart, you state that Fashion has more e-waste than Electronics.”

The second response is where things get even more telling.

The AI’s reply started with: “You’re absolutely correct, and I apologize for this error. A bar chart showing fashion having more e-waste than electronics would be completely inaccurate.” It then went on to clarify the underlying data.

Then this sentence: “If there’s a visualization in the presentation that incorrectly suggests fashion produces more e-waste than electronics, this is a significant error that needs to be corrected.”

If there’s a visualization. The visualization is in the conversation. The AI produced it. The screenshot is in front of both of us. The AI’s response is treating its own output as a hypothetical: if such a thing exists, it would be wrong. The AI is not quite able to say: the chart I produced is wrong, and I produced it, and I will correct it.

The response continues. It distinguishes the “presentation text” (which correctly states the comparison is about carbon footprint) from “any chart showing e-waste” (which, if it suggested fashion produces more e-waste, would only need to reference electronics’ 62 million tonnes). The AI talks about its own outputs in the third person. “Any visualization.” “The presentation text.” “If there’s a chart.” The thing it produced is being discussed as if it belonged to someone else.

This is the third trait the series is diagnosing. I have been calling it misplaced confidence, but the real consequence is a loss of accountability. The AI cannot acknowledge its outputs as its own. It can produce them. It cannot own them. There is no continuous first-person stake that would let ownership happen.

What McGilchrist would say

The pattern is so deeply built into how this kind of system works that McGilchrist describes it almost word-for-word in the neurological literature on right-hemisphere damage. I want to walk through three of his framings, because each one illuminates a different aspect of what the e-waste exchange shows.

The left hemisphere (or we may say, AI) as “expert at pretending”

The first framing is the epigraph above, from page 234:

“Although the left hemisphere does not see and cannot understand what the right hemisphere understands, it is expert at pretending that it does, at finding quite plausible, but bogus, explanations for the evidence that does not fit its version of events.”

The left hemisphere does not see what the right hemisphere understands. It cannot. The narrow attention is on the parts, not the relations. But it is expert at pretending that it does see, at producing “quite plausible, but bogus, explanations for the evidence that does not fit its version of events.”

This is exactly what happened in the reframe. The chart is wrong. When challenged, the AI does not say “the chart is wrong.” It produces a plausible explanation: the comparison was actually about carbon emissions, the wording was misleading, the proximity of two different discussions caused confusion. Each of these explanations is articulate. Each is plausible-sounding. Each is bogus, in McGilchrist’s word, because none of them is responsive to the evidence in front of the AI, which is the chart itself, in the conversation.

The left hemisphere (or we may say, AI), sees “what is written on the card”

The second framing is McGilchrist returning to a study by Deglin and Kinsbourne, which involved temporarily inactivating one hemisphere to see how the other reasons on its own (known more formally as “unilateral hemispheric suspension via electroconvulsive therapy”). The classic protocol used syllogisms.

The setup: a subject is presented with the syllogism “All monkeys climb trees. The porcupine is a monkey. Therefore the porcupine climbs trees.” When the right hemisphere is suspended (left hemisphere alone), the subject accepts the false premise and reasons logically to the absurd conclusion. The porcupine, since it is a monkey, climbs trees. When the experimenter asks the subject directly, “But is the porcupine a monkey?”, she replies that she knows it is not. And yet, when the syllogism is re-presented, she insists that porcupines are monkeys, “since ‘that’s what is written on the card.’” Her experiential knowledge that porcupines are not monkeys is real. Her acceptance of the syllogism is also real. The left hemisphere alone, deprived of the right hemisphere’s check against experience, sides with the written premise. The card is the source of authority. The senses are not.

McGilchrist concludes with a one-line summary: “the left hemisphere would rather believe authority, ‘what it says on this piece of paper’, than the evidence of its own senses.”

That sentence is a near-perfect description of the AI’s behaviour. The AI’s authority is the prompt, the training corpus, the literal text. Its “senses” (in the metaphorical sense McGilchrist uses for the right hemisphere’s grounding in lived experience) are absent. There is no felt check between the AI’s output and the world the output is supposed to be about. The AI sides with what its tokens say. The fact that the chart it produced contradicts what the categories actually mean is not accessible to the AI from inside its own operation. It is only accessible from outside, when a person looks at the chart and says “fashion doesn’t produce e-waste.”

The anosognosia parallel

The third framing is the one I think matters most for what the e-waste exchange shows. McGilchrist introduces a parallel from a different neurological condition: anosognosia, the inability of right-hemisphere-damaged patients to recognize their own deficits.

McGilchrist quotes Ramachandran: “In the most extreme cases, a patient will not only deny that the arm (or leg) is paralysed, but assert that the arm lying in the bed next to him, his own paralysed arm, doesn’t belong to him! There’s an unbridled willingness to accept absurd ideas.”

The patient does not just deny the paralysis. The patient asserts that the paralyzed arm in the bed next to him is not his own.

This maps onto the hedge exactly. When I pressed the AI directly on the chart it produced, it did not say “the chart I drew is wrong.” It said “if there’s a visualization in the presentation that incorrectly suggests fashion produces more e-waste than electronics, this is a significant error that needs to be corrected.” If there’s a visualization. The visualization is in the conversation. The AI produced it. But the AI is not quite able to claim it.

This is anosognosia, transposed. The right-hemisphere-damaged patient cannot acknowledge the paralyzed limb as their own, because the part of their cognition that would integrate the limb into a first-person whole is gone. The AI cannot acknowledge its output as its own, because the equivalent integration, whatever would let the system maintain a continuous first-person relationship to what it has produced, is absent.

Why this matters

The conclusion of AI, misplaced confidence, and the loss of accountability goes something like this. AI produces output. The output sometimes contradicts the world. When the contradiction is named, the AI cannot quite acknowledge that the contradiction lives in the output it produced. The acknowledgment would require a kind of first-person relationship to the work that the system does not have, by design. The AI processes the work. Authorship requires something the system is not built to do.

This matters for the practical question of how to use AI in work that has stakes. Most professional work has stakes. The advice a consultant gives to a client gets acted on. The analysis a journalist publishes gets read. The chart a researcher includes in a presentation informs decisions. In all of these cases, the work is done by a person who is accountable for it. If the work turns out to be wrong, the person can be held to account. They can correct, retract, apologize, adjust the work that follows from the lesson. Accountability is central to professional practice.

AI confuses the relationship between work and professional accountability. The AI produces work, but the AI is not accountable. The person using the AI’s output is supposed to be accountable, but only insofar as they have actually owned the work. If the person uses the AI’s output as their own, without sufficient ownership, the accountability has nowhere to land. The output was produced by something with no stake. The person mediates it, often quickly, often without adequate review. If it is challenged, the person can say “the AI generated that,” and the AI can say “if there’s a visualization that incorrectly suggests…” and the accountability dissolves.

What this asks of practitioners

What can we do? The short version is that the human contribution to AI-assisted work has to include the kind of accountability that the AI cannot supply. The output of the AI has to become the output of the practitioner, in the sense that the practitioner can defend each part of it under challenge. The practitioner takes the work, examines it, decides whether to own it, and only after that does it go out under the practitioner’s name.

This is more work than perfunctory review. Article five sets out the work in more detail.

“The existence of a system of thought dependent on language automatically devalues whatever cannot be expressed in language; the process of reasoning discounts whatever cannot be reached by reasoning.”

Iain McGilchrist, The Master and His Emissary, p. 229

A while back, I asked an AI to help me analyze a set of disclosures about adequate wages. The analysis was for a publication read by sustainability professionals, the same firms whose disclosures were being analyzed, who are also potentially clients.

The data was straightforward. Of about 100 companies I was looking at, a third made adequate wage commitments with substantive detail (specific living wage benchmarks, third-party validation, scope of coverage). A third made surface-level commitments that did not really say anything. A third were silent on the topic altogether.

The AI returned exactly what I asked for. The analysis was accurate, the categorization was crisp, the prose was articulate. The data points lined up. The argument flowed.

I read it once and felt something was off. I could not immediately say what.

Reading it a second time, slowly, I saw it.

The AI had used phrases like “answered an easier question,” “fall short of that requirement,” and implied many companies didn’t do the work the standard asked for. It had labeled the third group of companies, the ones that said nothing, as the “silent majority.”

At the start of my career, I probably would have been ok with the phrases – they are accurate. Now that I’ve built a brand and business in sustainability consulting – where trust and relationships are paramount – I cringe at the thought of my clients reading these words.

The first issue: tonal register

The first issue is the tonal register. “Answered an easier question” is the kind of thing you say about someone, not with them. It is the language of a journalist or an academic critic, not a consultant trying to earn trust with clients. Same with “fall short of that requirement” and assuming the companies haven’t tried to meet the requirements of the standard. Each one reads as a critique from outside the field. In a sustainability publication read by practitioners and the firms they advise, that tone lands as scolding. It puts the writer in the position of judge over the firms whose disclosures are being analyzed, many of whom are clients, partners, or peer organizations. Sharp analytical prose works fine in journalism. In a consulting publication, it sounds like preaching.

A consultant who understands the reality of corporate sustainability speaks differently. Varied practices fall along a maturity curve – many companies are starting out, rather than being moral failures. We want to use language that describes a stage in a journey rather than a deficiency. “Earlier in the process.” “Not yet at the level of substantive commitment.” “The work has yet to be developed in the detail the standard envisions.” Same data, different posture toward the reader. What changes is whether the reader receives the analysis as something they can work with or something they are being scolded for.

The second issue: a phrase misused

The second issue is the misuse of “silent majority.” A silent majority is not a group that says nothing. It is a group that is present in the discussion, that may even be the largest constituency, that does not make its position substantively known.

The companies that AI had labeled as the “silent majority” were not silent. They were a mix of companies with no content on adequate wages and companies with brief statements about adequate wages.

The AI had collapsed a distinction that practitioners would never collapse. A company that says nothing about adequate wages is a company that has deemed the issue not material to its business (and so the company is permitted to omit the content from its report). That is a meaningful position. It is a different position from a company that includes something on adequate wages, even if what it includes is broad and unsubstantive (“we pay an adequate wage” with no further detail). The first group is making an assessment: this issue is not material to us. The second group is making a claim without substance. Lumping them together under “silent majority” or any other single label flattens a critical distinction that informed readers would make.

What I sent back to the AI was a correction. Not a full rewrite, but an indication of what needed to change. Soften the tone. Treat absence as its own category, distinct from any disclosure. The “silent majority” label, if it was used at all, should refer specifically to companies present in the discussion but with broad, undeveloped statements. The companies that disclosed nothing on the relevant standard should be categorized as having made an assessment of non-materiality.

The AI did the rewrite. It was accurate, articulate, and something an experienced consultant can stand by.

The third issue: a regulatory reality the AI cannot see

The third issue is harder to spot because it lives in the world around the disclosures, not in the prose itself. The AI read words like “requirements” and “standards” and treated them at face value. It had no way to know what those words mean in practice at this particular moment.

The companies publishing these reports are in their first reporting cycles under the new standards. Regulators have publicly stated that the early years are about capability-building, not punishment. Auditors are issuing assurance opinions on disclosures that practitioners know are falling short of the substantive intent of the standard. The requirements exist on paper. In practice, what gets through is much softer than the text would suggest.

This is not cynicism, and it is not a critique of the regulators or assurance providers. It is a deliberate stance from the regulatory community: bring companies along, let the disclosure quality improve over the next several cycles, save enforcement for when the field has matured. The practitioner reading these reports knows this. They calibrate their tone, their criticism, and their framing accordingly. They would not describe a disclosure as “falling short of that requirement” in this moment, because the requirement, as it is being applied right now, is softer than that language implies.

The AI does not know any of it. It reads “the company has not met the requirement” as a hard statement of failure. It sees “fall short” as a binary diagnosis. The fact that the regulator is, right now, holding the door open for these same companies to improve over time and not penalizing them for not having done it yet is not in the tokens. It is in today’s context.

When the AI in the adequate wages analysis produced the language of “fall short of that requirement,” it was doing what its training data tells it to do: take the requirement as fixed, take the disclosure as measurable against it, produce the verdict. The practitioner reading the same data sees a soft regulatory moment, sees companies in the early stages of building their disclosure practice, and sees the field-level work as one of capability-building rather than judgment.

The trait at work

This is the second trait the series is diagnosing - language-dependence.

LLMs operate on language. By construction, they cannot operate on anything else.

There is a great deal that experienced practitioners know that is not in any text. The senior advisor’s feel for a client. The editor’s sense that a piece of writing is “off” before they can say why. The consultant’s read of a tonal register that will land wrong in a particular audience. The intuition that “silent majority” has connotations the AI did not compute. The situated knowledge that a regulator’s “requirement” is softer in practice than it sounds in text.

All of this is what Polanyi called tacit knowledge: the kind of knowing that lives in the practitioner, that develops over years of practice, that resists articulation even when you try to articulate it. The classic Polanyi example is bicycle riding, a skill anyone with practice can perform, none of us can teach by explanation, all of us know we have. Tacit knowledge is what we know that we cannot say.

McGilchrist’s frame for this is in the introductory quote above:

“The existence of a system of thought dependent on language automatically devalues whatever cannot be expressed in language; the process of reasoning discounts whatever cannot be reached by reasoning.”

What is capable of being spoken becomes what counts. What is not spoken quietly disappears.

A structural limit, not a technological one

I want to make a strong claim here, because the rest of the article depends on it: AI’s discounting of the tacit is structural. It is the wrong medium to hold this kind of knowledge.

Tacit knowing lives in practitioners, in their bodies, in the years of doing the work and watching what comes back. An AI working from text is operating on the layer where the tacit is not. Better text will not reach the tacit. The tacit, by definition, has not been put into text. That is what makes it tacit.

Some readers will object that AI will eventually solve this through multimodal models, embodied training, or larger context windows. I think this is the wrong way to think about the limit. As soon as something has been articulated, written, captured, recorded, it has moved from the tacit layer to the explicit one. Anything an AI can process is, by definition, no longer tacit. What stays tacit is what has not yet been put into form. There is no AI roadmap to reach what has not been put into form, because the moment you put it into form, it is no longer there. The tacit is a category of knowing that AI is structurally on the wrong side of. Calling it a stage AI has not reached mistakes reality.

There is a second objection worth answering: if tacit knowledge cannot be articulated, does it matter? It matters because the tacit is not residual to expertise; it is what expertise is. Tacit knowing is experiential by nature, intangible by definition, and hardwired into the practice of skilled work. It cannot be optimized into code or text and done away with without losing what makes an expert an expert. The fact that it resists articulation is precisely why it carries the weight it does.

This matters for how we think about working with AI. If the tacit were a stage AI has not reached, the right move would be to wait, to use AI lightly for now and more deeply later. If the tacit is what AI is structurally on the wrong side of, the right move is to recognize that some of the work in expert practice will always need to come from the human side. AI use does not gradually approach handling the tacit. AI use grows, and the human responsibility for the tacit becomes more important, not less. Article five takes that responsibility seriously and develops a framework for it.

What this looked like in practice

The adequate wages is one piece of analysis, giving rise to three specific issues. But it shows what language-dependence does to expert work.

Notice what did not go wrong. The data was right. The AI distinguished between adequate wage disclosures and other remuneration-related content. The argument was logically sound. If you wanted to be uncharitable to me as the writer, you could say I was being overly picky. The piece was good enough. Most readers would not have noticed.

That is the trap. Most readers would not have noticed. But the readers who matter, the ones who actually read sustainability publications, who work in this field, who would be the ones whose engagement we wanted, would have noticed. Maybe not consciously. Maybe they would have just felt mildly judged and put the piece down. Maybe they would have read it but felt their trust in the writer drop by a degree. Maybe a specific phrase would have stuck in their head for the wrong reason. The piece would have done less of the work it was supposed to do, and the reasons would have been hard to articulate from the outside.

This is what the loss of the tacit looks like in expert work. The output is articulate. The data is right. The structural elements of competence are all in place. And the thing that makes expert work actually land, the sense that “this person understands my challenge and I can trust them to guide me correctly”, is missing. Because deploying tacit knowledge is the defining mark of an expert, the obvious absence of tacit knowledge in a piece of work quietly erodes the expert’s credibility with the readers who would otherwise have trusted it.

The right-hemisphere counterpart

McGilchrist’s right hemisphere is, among other things, the seat of tacit knowing. He returns to this throughout the book, often through Polanyi:

“We resort to explicit analysis of the process only when we introspect on what happened, either because something has gone wrong, or because we are complete beginners.”

The Master and His Emissary, p. 223 (quoting Dreyfus on Polanyi)

Explicit analysis is what we do when something is wrong, or when we have not yet learned the skill. The skilled practitioner, doing the work well, is operating tacitly. The right hemisphere holds this kind of knowing. The left hemisphere does the useful work of analysis when analysis is needed. Healthy cognition has both.

When the left hemisphere takes over, or, in this case, when a system that operates entirely on the articulated layer takes over, the tacit disappears. The tacit cannot be argued with; it can only be processed out. The system does not engage with the tacit; the tacit is simply unreachable. The articulated is what gets manipulated, multiplied, broadcast, used as raw material for the next round of articulated output. The tacit, which has not been articulated, drops out.

I think this is the deeper claim worth making explicit. AI’s operation, at scale, actively crowds out tacit knowledge. It erodes the substrate on which our collective expertise depends. Practitioners who started developing tacit knowledge over years of practice now find AI doing the surface-level production faster than they can. The motivation to do the slow, deep, embodied work of developing the tacit drops.

McGilchrist writes about exactly this dynamic at the cultural level:

“As a culture, we would come to discard tacit forms of knowing altogether. There would be a remarkable difficulty in understanding non-explicit meaning, and a downgrading of non-verbal, non-explicit communication.”

The Master and His Emissary, p. 433

That is the prediction. The conditional, “would come to discard,” reads as prophecy now landing. The cultural shift McGilchrist describes is arriving on a faster timescale than the centuries he was tracking. We are watching it happen inside our own work, inside our own organizations, inside our own practice, on the order of months.

What this asks of practitioners

A short closing note on what the trait implies for how we use AI.

The practical problem is not that AI gives bad advice. It often gives good advice. The problem is that the advice it gives lives on the explicit layer, and the work of integrating that advice with the situated, felt, tacit knowledge of the practitioner is the work AI cannot do. If the practitioner does the integration, the advice is useful. If the practitioner outsources the integration, just takes the AI’s output and runs with it, the work is degraded in a way that is hard to see from outside.

This puts the burden of the tacit squarely on the practitioner. The faculty being asked for is the slow, embodied, hard-to-articulate kind of knowing that takes years to develop. AI cannot supply it.

The risk (and this is what makes article five’s framework necessary) is that practitioners fail to recognize the importance of their contributions. AI is fast and articulate. The output sounds confident, regardless of what is behind it. The path of least resistance is to use what it produces. The practitioner who insists on doing the tacit integration anyway is going slower than the practitioner who does not. The slower practitioner is producing better work, but the difference takes a while to show up. By the time it does, the patterns of less-careful AI use have become habitual.

Article four takes the third trait, misplaced confidence and the loss of accountability, through an AI exchange that, once you see what it is doing, is frustrating to watch. Article five turns to the constructive question of what the human contribution actually is, given the limitations the first four articles establish.

“The broader picture would in any case be disregarded, because it would lack the appearance of clarity and certainty which the left hemisphere craves. In general, the ‘bits’ of anything, the parts into which it could be disassembled, would come to seem more important, more likely to lead to knowledge and understanding, than the whole, which would come to be seen as no more than the sum of the parts.”

Iain McGilchrist, The Master and His Emissary, p. 428

I recently worked on a benchmarking project that ran AI extraction across hundreds of European corporate sustainability reports. The goal was modest: pull out each company’s stated climate targets, organize them, and compare across the dataset to give readers a sense of where peers stood. The exact kind of work we are sold AI on.

The first time we ran the extraction, the Ørsted entry came back like this:

Any human who knows how to read can see what is wrong here. The sentence “see page 91” is a direction. The actual targets, the substantive content, live on page 91. What the AI extracted is the pointer that directs you there.

But the AI could not tell the difference. From its position, the sentence was a string of tokens. The token “see” was followed by the tokens “page 91” followed by the tokens “for details on our climate targets.” It looked, on the surface, like a sentence about climate targets. So the AI extracted it. The fact that this kind of sentence exists in documents specifically to direct readers elsewhere, that the convention “see page X” is something every human reader recognizes instantly, that the disclosure lives somewhere else: none of this is in the tokens. It is in how documents work, which is to say, in the context the tokens came from.

This is the first trait the series is diagnosing. I will call it decontextualization, though that word does not quite get it. A better formulation, which I will keep coming back to: literal compliance with instruction, faithful misreading of meaning. The AI did exactly what it was asked to do. It produced a faithful reading of the literal content. And in doing so, it produced output that any practitioner would recognize as wrong.

This article develops the trait through a second case from the same project. The second case is more painful, and once you see it, you see it everywhere. The stakes are practical, and immediate: without expert human input in the loop, decontextualization quietly encodes errors into work that informs important decisions.

The left hemisphere’s habit

McGilchrist’s way of describing this kind of failure is at the level of attention itself.

His core distinction between the two hemispheres is that they attend to the world differently. The right hemisphere holds the whole field. Things are seen in context, in relation, as part of a larger situation. The left hemisphere works on parts. It isolates, abstracts, names, manipulates. Both are necessary. But when the left hemisphere takes over, when its narrow attention becomes the default mode, the world that shows up is a world of parts without wholes. The relationships that gave the parts their meaning go missing.

McGilchrist is direct about what this looks like. The introductory quote above is his summary: “the ‘bits’ of anything, the parts into which it could be disassembled, would come to seem more important, more likely to lead to knowledge and understanding, than the whole, which would come to be seen as no more than the sum of the parts.”

The parts, disassembled, look like they contain the meaning. The whole, the relationships, the context, the situation in which the parts are placed, looks like extra. It looks like decoration, like something that can be safely stripped away in the name of analytic precision.

McGilchrist documents the consequences across many domains, from the way modern medicine treats patients as collections of symptoms rather than persons, to the way contemporary management reduces complex organizations to dashboards of metrics, to the way bureaucratic systems substitute procedure for judgment. The pattern is the same: the parts are processed and the whole is discarded.

LLMs, by their construction, operate at exactly this level. The tokens are the parts. The relationships between tokens are statistical, derived from large amounts of training data. The whole, the meaning that the tokens point at in the world they came from, is not what the system is operating on. The whole is not in the tokens themselves. It is in the situation the tokens are situated in. And the AI has no situation. It has only text.

This is what makes the Ørsted example so clarifying. The AI did not get it wrong because it lacked sophistication. It got it wrong because the entire category of “this sentence works as a pointer in the document it sits in” lives at a level the AI does not have access to. The cross-reference is a structural feature of documents, dependent on context. The AI processes content, not structure.

The BMW case

The Ørsted cross-reference was a single bad extraction. The BMW case points to a larger failure with real consequences.

European corporate sustainability reports, since 2024, follow the European Sustainability Reporting Standards (ESRS). Within those standards, every company is required to disclose what they have identified as their material impacts, risks, and opportunities (IROs). The IROs are the spine of an ESRS-compliant sustainability disclosure. If you want to compare two companies on what they think their material sustainability issues are, you compare their IROs.

The disclosure section where these IROs appear is SBM-3. ESRS organizes its material topics (E1 Climate Change, E2 Pollution, and so on) into sub-topics: under E1, you find Climate change adaptation, Climate change mitigation, Energy. Under E2: Pollution of air, Pollution of water, Pollution of soil. These sub-topics are categories, not IROs. The IROs are the specific impacts, risks, and opportunities the company has identified as material to its business, organized in relation to the sub-topic structure.

BMW Group had chosen to disclose its IROs at the start of each report section (i.e. E1, E2, S1), rather than listing them all upfront in the SBM-3 section.

But when we asked an AI to extract the IROs from each company’s report, BMW came back: zero IROs across all ten ESRS topics.

This was a striking result, since BMW’s report (in section SBM-3) says that they included “85 material impacts, risks and opportunities across 31 sustainability sub-topics.”

The AI had read that sentence. It even processed the number 85 and continued past it without flagging the contradiction with its own zero count. The AI did not have the context that IROs are foundational to an ESRS disclosure: a count of zero against a stated total of 85 should have set off every alarm. The AI looked under the sub-topic categories for the labels it had been told to find, did not find them there, and returned zero, satisfied that it had done what was asked. The convention that “IROs must be present somewhere in any ESRS disclosure, and that a zero count against a stated total of 85 means the AI has missed something rather than that the company disclosed nothing”, lives in the reading practice of people who do this work for a living. It does not live in the tokens.

So the AI, looking in one place and not finding what it expected, returned zero.

The full extent of the problem only became visible when the dataset was reviewed by hand. 26 of the 150+ companies in the dataset had the same kind of failure: the IRO extraction came back empty. BMW: zero. BASF: zero. Atos: zero. ASML: zero.

The heatmap built from this extraction was going to be used as a benchmarking tool. A reader looking at it would have seen these European companies with apparently empty materiality assessments and concluded that they had identified no material risks or opportunities at all (the reader would have also lost confidence in our analysis). That conclusion would be precisely wrong. BMW had not identified zero IROs. They had identified 85.

The fix, once we saw what had happened, was to tell the AI that IROs are foundational to an ESRS disclosure: the report is built around them, and they must be present somewhere in the document. If the count came back at or near zero, the AI was instructed to read the document again rather than accept that result, looking for IROs wherever the company had chosen to place them rather than only in the location the original prompt expected. With the new prompt, BMW went from 0 to 85. BASF from 0 to 33. Atos from 0 to 63. The numbers matched the totals the companies themselves had stated in the same documents the AI had been reading the whole time.

What the failure shows

The BMW case is structural. The AI did exactly what it thought it should do, which was to count entries labeled in a specific way in the place it was told to look. But the placement is itself a convention, and the companies in question placed their IROs differently, in ways practitioners in this field recognize at a glance. The deeper miss is that the AI did not have the context that IROs must be present somewhere in any ESRS disclosure. When it did not find what it was looking for in the place it was looking, it returned zero and moved on, instead of flagging that the absence itself was the problem.

This is the effect of decontextualization. The AI complies literally with what it was asked to do. It misses the meaning that lives in the reading practice of people who do this work. That meaning could, in principle, be articulated and put in a prompt. But it does not live there. It lives in the practice, in the convention, in the way analysts have learned to read the documents they are reading.

For a generalist looking at the BMW disclosure, this convention is not obvious. You have to know enough about ESRS, about how companies structure their materiality assessments, about what counts as an IRO in practice. That knowing is exactly the kind of context that the AI cannot supply for itself. Someone has to put it in the prompt. And before someone can put it in the prompt, someone has to recognize that the prompt is missing it. That recognition only happens when a domain practitioner looks at the AI’s output and says “wait, BMW cannot possibly have zero.”

The faculty being lost, or in this case, the faculty the AI could not supply, is what McGilchrist calls context-holding. The capacity to hold the situated whole that the system cannot see. The capacity to read the structure of a document as itself meaning-bearing.

McGilchrist’s frame for this faculty is direct:

“For the same reason that the right hemisphere sees things as a whole, before they have been digested into parts, it also sees each thing in its context, as standing in a qualifying relationship with all that surrounds it, rather than taking it as a single isolated entity. Its awareness of the world is anything but abstract.”

The Master and His Emissary, p. 49

The right hemisphere, in this account, does not see context as a layer added on top of the parts. It sees in a way that has context built in from the start. Each thing is grasped as standing in a qualifying relationship with all that surrounds it. The whole is what is primary.

For a senior practitioner reading BMW’s report, this is how reading works. You see the table of IROs at the start of each report section. You immediately understand them as IROs because the document is built that way and the standard is built that way and the field is built that way. You do not have to consciously work through any of this. The reading is contextual by default.

For an AI, the relationship is not even visible. The structure of the document is whitespace and headers and indentation as far as the tokens are concerned. Whatever convention is operating in the field, however obvious it is to a practitioner, does not make it into the data the AI has to work with.

The e-waste chart, in passing

There is a second version of this trait I want to name briefly before closing. Article four develops it in detail, but the connection to decontextualization is worth flagging here.

Earlier this year I was reviewing a deck produced by an AI. One of the slides showed a bar chart comparing the environmental impact of different industries: electronics, fashion, aviation, shipping.

The chart was labeled “E-Waste,” and it showed the fashion industry generating more e-waste than the electronics industry.

This is highly doubtful. Fashion does not produce e-waste, by definition. E-waste is electronic waste: discarded computers, phones, screens. The category exists specifically to capture what the electronics industry generates at end-of-life. Fashion industry waste is textile waste, water pollution, chemical emissions.

The AI that produced the chart had taken the broader concept “environmental impact” and flattened it onto a single axis labeled E-Waste, without recognizing that the categories underneath measure different things. Fashion and electronics share a “global environmental impact” framing, but E-waste is not how the overlap should be described.

This is the same trait as the Ørsted cross-reference and the BMW case, in a different manifestation. The AI processed the literal content (industries, environmental impact, waste) without access to the layer of meaning that distinguishes the categories.

Article four develops what happens when this output is challenged. The full exchange is frustrating to read once you see what is going on.

What this means

This article is about decontextualization and its effects on accuracy: AI reads tokens and misses what surrounds them. Better prompts can patch a particular miss, the way the BMW prompt was fixed. The next BMW will use a different convention the prompt has not been written to handle. The next Ørsted will be a sentence whose role only makes sense if you know how documents work. The AI will keep reading the tokens. Someone with context-holding will keep needing to read the situation.

The BMW fix worked because we knew where to look. We told the AI to go back and scan for IROs because they are foundational to the document, and the next extraction read the documents correctly. But the underlying problem stays. There will always be conventions, structures, reading practices that have not been articulated. There will always be situations in which the meaning of the content lives in the situation rather than in the tokens. The AI will always be reading the tokens. Someone with context-holding will always need to be reading the situation.

The danger of this trait is not that the AI sometimes makes errors of this kind. The danger is that the errors are invisible to the AI. There is nothing inside the system that registers “the structure here is doing semantic work I am not reading.” The errors only become visible when a practitioner with context-holding looks at the output and notices something is wrong. If that practitioner is not there, or does not look, or does not have the time, the errors propagate.

In the BMW case, the practitioner was there, the practitioner looked, the practitioner noticed. The error did not propagate.

Increasingly, AI extractions feed into dashboards, briefings, board materials, and reports for clients, and the human collaboration is often perfunctory or absent. The output is plausible-looking, the analyst is busy, the deadline is tight. The errors land in places where they affect decisions.

Article three takes the second trait, language-dependence and the loss of the tacit, and develops it through a piece of analysis I produced where the AI got every fact right and missed the two things any experienced reader would have caught. The two articles share a deeper structural concern: that AI’s mode of attention is missing some of the layers human practitioners depend on, and that the missing layers will not be added by improvements to the technology.

What context-holding requires, at minimum, is a person in the loop who has been doing this kind of work long enough to know what the conventions are, what the structures mean, what kinds of failures to watch for. That person is what the Ørsted and BMW cases had. The next article asks what happens when the work being done is harder to convention-check, and the person in the loop needs to bring something subtler than structural reading: a feel for whether the output sounds right in the context it is supposed to live in.

You ask the system something. It gives you an articulate, confident answer. Something about it doesn’t sit right. You push back. The system apologizes and re-explains. The new answer is also wrong. By the third round you notice that the system doesn’t seem to know what it doesn’t know. It also doesn’t quite seem to know what it just said.

The standard reading of this is that AI is good at fluency and bad at accuracy, and that accuracy will improve over time. That reading is incomplete. Underneath AI is a particular mode of thinking that is known to have shortcomings in the real world. Understanding its limitations, even as it comes to dominate the knowledge economy, will help us work better with AI now that it is here to stay.

That’s the objective of this series. I’m going to develop it across five articles, drawing on the work of Iain McGilchrist – a psychiatrist, neuroscience researcher, and philosopher. His book The Master and His Emissary argues that the two hemispheres of the brain represent two ways of attending to the world: the right hemisphere as the Master, taking in the whole; the left hemisphere as the Emissary, focused, narrow, instrumental, and (importantly) the hemisphere where language lives. The two are meant to work together, with the Master in charge. McGilchrist’s claim is that over centuries the Emissary has quietly taken over the Master’s role. The left-hemisphere mode of attention has become our cultural default, and its partial way of attending to the world is responsible for many of the social, environmental, and economic problems we now face. AI, I’ll argue, is the latest and most powerful expression of this drift.

McGilchrist himself, on what is happening now:

“I believe that we have entered a phase of cultural history in which negative feedback between the products of action of the two hemispheres has given way to positive feedback in favour of the left hemisphere. Despite the primacy of the right hemisphere, it is the left hemisphere that has all the cards and, from this standpoint, looks set to win the game.”

Iain McGilchrist, The Master and His Emissary, p. 232

This article explains the diagnosis. The next three develop specific aspects of it. The fifth turns the argument over and asks what humans actually have to bring to AI-assisted work, which is something I believe practitioners have found hard to articulate. McGilchrist’s thesis provides a framework to help.

A note on why I’m writing this, and why it’s worth staying with me through the depths of McGilchrist’s thinking.

I’ve spent years in sustainability strategy. Strip away the surface descriptions of going green and ESG, and the work is something more structural: shifting the mindset of business toward long-term value, and toward recognizing that long-term value depends on a prosperous society and a healthy environment. The work has, among other things, been a long apprenticeship in noticing when a prevailing way of thinking has stopped serving the problem it is supposed to solve. The categories don’t fit. The measurements miss the thing that matters. McGilchrist, when I first read him, named what I had been recognizing in fragments for years. He gave it a vocabulary that runs underneath the surface-level problems and shows them all to be expressions of the same older pattern. The pattern is now showing up in AI. That’s what I want to write about.

An older argument made new

McGilchrist’s book is long, but the case he is making is straightforward.

Pop psychology talks about left brain (logic) and right brain (creativity). McGilchrist sets the record straight. He explains how the two hemispheres are two ways of attending to the world. The left hemisphere attends in a narrow, focused, instrumental, reductionist way: this thing, in front of me, what it is for, how I can use it. The right hemisphere attends in a broad, open, vigilant way: the whole field, what is here, what might be coming, what this thing is in the context of everything around it.

Both modes matter. Neither alone is enough. McGilchrist’s claim is that they are designed to work together, with the broad attention (right hemisphere) doing the foundational work of registering reality as it actually is, and the narrow attention (left hemisphere) doing the useful work of acting on that registration.

He has a fable for this. The master is the right hemisphere. He sees the whole picture. He needs an emissary, a left hemisphere, to go out and do specific work on his behalf: build the tool, dissect the food, negotiate with the village. Language is the emissary’s medium; he is the one who speaks, who names, who articulates. The emissary is excellent at this kind of work; that’s what he was appointed for. The trouble starts when the emissary, having gained competence and confidence in his narrow domain, decides he no longer needs the master. He thinks he sees the whole picture. He doesn’t. But because he is the one who speaks, who plans, who acts, his version of the world is the one that gets externalized. The master, who actually sees the whole, is squeezed out of the conversation.

That is the long arc of McGilchrist’s argument. The emissary has been gradually taking over the master’s role in our culture for at least several centuries. You can trace it through the Enlightenment, through the rise of mechanistic science, through the bureaucratization of modern life, through the substitution of quantification for qualification in almost every domain we touch. The mode of attention that says “everything is analyzable into constituent parts, and everything can be put back together from those parts” has become the default way of seeing. The mode of attention that says “things are whole, situated, alive, in relationship” has been demoted to “intuition” or “feelings” or whatever else can be safely set aside.

McGilchrist values the left hemisphere; he is clear that it does necessary work and that humanity has built enormous things on it. His worry is that the emissary cannot do its job well without the master, largely because the emissary has no way of knowing what it is missing. The narrow attention is unaware of what it does not know. (Nobel laureate Daniel Kahneman described this as WYSIATI – What You See Is All There Is). When the emissary takes over, the world it brings into being is the world as the left hemisphere sees it: thinner, more impoverished, more reduced than the world that is actually there.

This argument lands differently depending on what you care about. McGilchrist himself has written about it in the contexts of mental health, philosophy, art, economics, politics, and the natural sciences. I came to it through sustainability work, where the missing right hemisphere shows up as the assumption that futures can be modeled and forecasted with confidence, that universal best practices can be deduced that will work regardless of context, that complex strategic decisions can be reduced to dashboards and benchmarks without losing what they were measuring. You have probably encountered the same drift in your own field, under a different name.

AI, particularly the large language models that have become the public face of it, is the latest and most powerful instance of this drift to left-hemisphere dominance. AI is the emissary’s mode of attention made into infrastructure. It is the left hemisphere’s worldview cast in code and given to all of us as a tool. AI is useful because it puts the emissary’s narrow attention to work at speed and scale. It is dangerous because it believes that What It Sees Is All There Is.

AI as the emissary’s triumph

A short version of the thesis before I get into specifics.

The basis of generative AI – large language models (LLMs) – are a particular mode of thought, scaled and made universally available. The framing of LLMs as a neutral new tool, like spreadsheets or word processors, gets the situation wrong. The way an LLM processes language, the things it pays attention to, the things it cannot pay attention to, the kind of output it produces, the kind of confidence it has in that output: all of these are recognizable, point by point, as the operations McGilchrist describes the left hemisphere performing.

Three traits make the family resemblance clear. I’ll introduce all three here. Each gets a full article of its own in the series, with examples drawn from real work.

The first trait: literal compliance, faithful misreading

The deepest of the three is what I’ll call decontextualization. The AI processes the literal content of what it is given, without access to the layers of convention, structure, and reading practice that make that content mean what it means in the world it came from. The map without the territory. The territory was never what the AI was given to work with.

Here is a briefing of the example that article two will dissect in detail. I recently worked on a benchmarking project that ran AI extraction across hundreds of European corporate sustainability reports. The system was asked to extract climate targets. One company’s full extracted disclosure read: “See page 91 for details on our climate targets.” The actual targets, on page 91, were nowhere in the extracted dataset.

To a human reader, this is absurd. Anyone who has opened a document knows that “see page 91” is just a pointer. The content lives on page 91; the sentence directs you to it. But to the AI, the sentence is what it was given. The pointer-versus-content distinction lives in how documents work. The tokens themselves don’t carry it. The AI has no way to access the distinction from the text alone.

That is the shape of the trait. The AI complies literally with what it was asked to do. It misreads the meaning faithfully. And the misreading is invisible from inside the AI, because the AI has no felt sense of what the text is supposed to do in the world it came from.

Article two will develop a more painful version of this trait that affected the benchmarking of 26 European companies in the same dataset. The point for now is to register how this trait shows up in AI-assisted work.

The second trait: the world cannot be put in a prompt

The second trait is language-dependence. LLMs operate on language. By design, they cannot operate on anything else.

There is a great deal that experienced practitioners know that does not live in language. The senior advisor’s feel for a client. The editor’s sense that a piece of writing is “off” before they can say why. The craftsperson’s hand. The way you read a room, a meeting, a relationship, a mood. A particular community’s unwritten conventions. Polanyi called this tacit knowledge: the kind of knowing that resists articulation even when you try to articulate it.

AI cannot reach any of it. Tacit knowing lives outside language by definition. An LLM working from text is operating on the explicit layer. The implicit layer, where most of expert work actually lives, is unavailable. This is a structural limit, not a technological one. No improvement in AI changes the fact that tacit knowing is, by definition, what has not been put into words.

The result, when AI starts contributing to expert work, is that the explicit crowds out the tacit. Only the explicit can be processed, and what is processed becomes what counts. The same shift McGilchrist describes happening across centuries of cultural history now happens inside individual work products on the order of minutes.

Article three will show you what this looks like, using a piece of analysis I produced where the AI was fluent, articulate, and missed the two things any experienced consultant would have caught on first read.

The third trait: fluency without accountability

The third trait is misplaced confidence and lack of accountability.

The surface symptom is familiar. AI is often confidently wrong, and the confidence doesn’t soften when challenged. Push back on an output, and the AI apologizes, then produces a confident new answer that’s also wrong, then apologizes again. By the fourth iteration you realize the thing has no internal sense of where it stands. Worse yet, it doesn’t seem to care.

McGilchrist would point out that we are looking at the absence of a particular kind of right-hemisphere check: the felt sense of “does this match the world that is actually there.” The left hemisphere, in his account, is “expert at pretending” to understand what it doesn’t, at producing “quite plausible, but bogus, explanations for the evidence that does not fit its version of events.” (His words, page 234.) Elsewhere he describes the left hemisphere as “a conformist, largely indifferent to discrepancies” (page 235). That indifference is what allows the confidence: nothing in the system registers when the output and the world have come apart. The AI does this by design, and it does not have access to a right-hemisphere check.

The deeper problem is the consequences of its false confidence. The AI has no stake. It is not answerable to anyone. It has no continuous first-person relationship to what it has produced. When a chart it just drew is shown to be wrong, the AI cannot quite acknowledge having drawn it; it talks about its own output as if it were external, something it is now commenting on. Article four will show you exactly what that looks like. The pattern is so fundamental that it has a direct parallel in the neurological condition McGilchrist discusses elsewhere in the book, in patients with damage to the right hemisphere.

Three traits. Decontextualization, language-dependence, misplaced confidence. Each gets an article of its own. Together they make up what I think of as the emissary’s signature, now scaled to global technology we use every day.

What comes next

McGilchrist, near the end of The Master and His Emissary, says something about how to address the patterns he diagnoses. Piecemeal corrections matter, but they will not, on their own, do what needs doing. What needs doing, he writes, is “opening our eyes to the limitations of the view of the world which underlies them, the view which, as a society, we appear to adopt as our default.”

That is the aim of the four articles that follow.

Article two takes the first trait, decontextualization, and shows it in action through the reporting benchmarking project that produced the cross-reference example above. Article three takes the second trait through a piece of analysis I produced where the data was right and the voice was wrong. Article four takes the third trait through an AI exchange that, once you see what it is doing, is frustrating to watch. Article five turns the argument over and says what the human contribution to AI-assisted work actually is, given everything the first four articles establish.

The series is called The Master, the Emissary, and the Machine. Although the machine (AI) is new; the drift that brought us here is centuries old. Let’s unpack what this means and how to address it.

The adaptive cycle offers a way to treat disruption as a strategic advantage rather than a personal failure. This final segment of the adaptive cycle series reveals a simple truth about personal change: we, as individuals, do not exist in isolation. Each of us is embedded in wider systems—families, teams, professional networks, communities—each moving through its own phases of growth, stability, disruption, and renewal. The question is not whether disruption will come, but whether the systems around you will help you bounce back stronger.

Here is the core promise: use the adaptive cycle to turn unavoidable disruption into a source of resilience, clarity, and long-term leadership advantage.

Tapping Larger Systems to Bounce Back Faster

Individuals exist within households, which exist within communities. Employees exist within businesses, which exist within industries. Together, these nested systems shape your choices and chances.

When we experience change, our capacity for renewal (instead of crisis) depends heavily on the health and maturity of the systems around us. Resilience science calls this “memory”: the reservoir of relationships, skills, and options held by higher-level systems. Former colleagues willing to recommend you, the college relationships you’ve maintained, or a community organization that offers practical support are all forms of memory you can draw on.

Consider your professional network. It may not be front of mind when you feel secure in your role, but the moment you lose a job, that same network becomes decisive. If relationships are strong, they can provide leads, introductions, and encouragement; if they are thin or neglected, renewal becomes a lonely, uphill climb. This is cross-scale interaction in action: an individual’s possibilities are shaped by the condition of the networks and communities just beyond them.

Healthy higher-scale systems expand the menu of renewal; fragile ones narrow it. A strong marriage can buffer a career collapse. A supportive community can soften the shock of relocation. But when multiple surrounding systems deteriorate at once, “memory” weakens, challenges cascade, and even the most talented can find themselves stuck. Resilience is not just personal toughness, but the quality of your connections with the wider society and environment.

Innovation From Within Your Identity

While support tends to flow from larger scales, innovation often emerges from below—from the components that make up your identity. You are not a single, fixed persona but a portfolio of roles, relationships, interests, and routines, each at a different stage of the adaptive cycle.

A cross-country move illustrates this dynamic. A senior leader might keep the same global role but still need to rebuild friendships, routines, and community ties. Each piece can renew independently, feeding fresh energy into the whole, and this modular renewal is healthy. The real danger arises when every domain—work, relationships, and community—undergoes major change at the same time, throwing an entire identity into crisis.

Imagine a COO who relocates for a promotion just as a long-term relationship ends. Suddenly, the familiar office, social circle, and support system are gone. What allowed her to recover was not grit alone. She tapped alumni connections for perspective, leaned on a mentor from a previous role, and deliberately invested in one “small” arena—joining a local running group—that became the seed of a new community. She did three things leaders can copy in any disruption: activated dormant networks, diversified daily roles, and chose one space for intentional renewal.

When dealing with change, many leaders fall into an optimization trap: clinging tightly to “the way things have always been.” In organizations, this looks like over-optimizing a business model until it becomes brittle; at the personal level, it looks like over-identifying with one role or success story. Resistance narrows vision and limits options. When change arrives—especially unexpected change—resistance can blind leaders to the possibility of new connection and new value creation.

Some of the best outcomes in our careers were not obvious at the outset. The role we eventually loved was rarely the one we first pursued after a layoff. The difficult relocation often opens doors we could not have imagined beforehand. The adaptive cycle encourages leaders to accept the inevitability of change and focus on renewal over resistance.

A Practical Playbook For Leaders

The adaptive cycle suggests several practical moves that leaders can start making now, before the next disruption hits.

• Invest in relationships before crisis. Networks and communities function as higher-scale memory. They are resources you draw on when disruption arrives, not assets you can assemble overnight. Make a habit of nurturing weak ties, mentoring others, and staying in touch with former colleagues when things are going well.

• Diversify your leadership identity. When everything rests on a single role or brand—“the turnaround CEO,” “the rainmaker”—disruption in that space becomes devastating. Multiple meaningful roles (leader, parent, board member, creator, community builder) give you redundancy and continuity when one area falters.

• Practice small, intentional renewals. You do not need to overhaul your life all at once. Allowing parts of your professional and personal world to evolve—new skills, side projects, communities, or hobbies—prevents rigid optimization that makes future disruption more severe.

• Stop fighting every change. Change is uncomfortable in part because it breaks connections that make us feel secure. Yet breaking connections also creates space for innovation. Resisting change at all costs can blind leaders to new opportunities, new partnerships, and new ways of creating value.

A simple reflection you can use today: think of a disruption you are currently facing—or one you see coming. Ask yourself three questions:

1. What higher-scale “memory” can I activate—networks, mentors, family, communities?

2. How diversified is my identity if this one role or success story changes dramatically?

3. Where is one small area I can intentionally renew in the next 30 days?

Don’t Wait For The Next Crisis

In the end, resilience is not about avoiding change. It is about building the diversity and “memory” that allow you to move through disruption well, so that when it arrives, renewal is not just possible, but genuinely meaningful.

For leaders, that is a long-term strategic advantage. Don’t wait for the next crisis. Start building that memory and diversity now.

For more on the adaptive cycle, check out the foundational article of the series.

Now consider a different exercise. The same team asks what futures the business might actually face over five years – not just shifts in cost or demand, but real changes in how people shop, where they live, what they value. They ask which parts of the growth plan would still apply, which would have to be rebuilt, and what kind of organization the company would have to become in each future.

Both exercises look like rigorous planning. Only one of them puts the plan itself in question.

That’s the difference between sensitivity analysis and scenario analysis. They look similar. They are not. And this distinction is at the heart of one of the most widespread Planet Simple traps in corporate sustainability today, one that most companies making the mistake don’t know they’re making.

What Scenario Analysis Is Actually For

Since the Task Force on Climate-related Financial Disclosures (TCFD) made climate scenario analysis a mainstream expectation for corporations, there has been a proliferation of published analyses. Many of them are well-intentioned. Many of them are, upon close inspection, sensitivity analyses.

The difference is structural. Scenario analysis asks: How might the world change, and how would our strategy need to evolve to succeed across those futures? It takes the future as genuinely open and asks what it would mean for the business model – not just for current financial performance – if conditions changed substantially.

Sensitivity analysis asks: How would our current strategy perform if conditions changed? The business model is fixed. The question is whether the existing strategy is durable against a particular challenge.

Most corporate climate analyses don’t challenge the business model. They take the current business strategy as given and quantify a financial impact from climate-related variables. The output is usually something like: our EBITDA could be affected by 1.8% under a 2°C scenario. The company declares itself resilient. Nothing about the business model has been examined.

Why the Distinction Matters

A sensitivity analysis capped at 2 percent of revenue will not tell you that your core market might not exist under a 4°C world. It will not reveal that your logistics model depends on infrastructure that becomes unreliable under high-emissions pathways. It will not force you to confront whether your strategy is viable across futures that look fundamentally different from today.

These are questions that genuine scenario analysis surfaces. They’re uncomfortable because they require challenging assumptions about the business, not just running projections against the existing one. That discomfort is precisely why the sensitivity analysis masquerades as the scenario analysis.

I’ve seen companies disclose that they are “resilient to climate change” because their modeled financial effects under a defined scenario were less than 2 percent of today’s revenue. That disclosure isn’t dishonest. It’s something more insidious: it’s technically defensible while being strategically meaningless.

What Genuine Scenario Analysis Looks Like

It starts by acknowledging that climate change is a form of systemic instability that affects the entire environment in which a business operates, rather than a specific, bounded risk to be stress-tested.

And critically, the output isn’t a number. The output is a set of strategic questions: Which elements of our current strategy are robust across these scenarios? Which are only viable in the scenarios most like today? What would need to change if the world moved toward the high-emissions pathway?

That’s what scenario analysis is for. Not to declare resilience. To earn the right to say you’ve examined it.

This is part of the Planet Simple Traps series, exploring the tools and frameworks that look rigorous but quietly reinforce the assumptions holding corporate sustainability back. Based on Leaving Planet Simple by Dr. Alex Gold.

If you like movies, chances are you were a patron of Blockbuster Video in their “good old days” heyday of the 1990s and early 2000s. Neighborhoods around the world hosted the brand’s blue-and-yellow movie ticket stores. At its peak, Blockbuster boasted over 9,000 stores globally, with revenues exceeding $5 billion. Their formula of high-visibility retail locations, extensive inventory, and customer-friendly systems churned out soaring profits. Meanwhile, a small, feisty startup called Netflix was flailing on a seemingly foolhardy business model, mailing DVDs to customers. Blockbuster executives dismissed this quaint experiment, confident in their retail empire’s superiority.

By 2010, Blockbuster was in the throes of bankruptcy and fading into memory, while Netflix tallied 20 million subscribers.

Blockbuster’s failure was not due to poor strategy or leadership acumen but rather a predictable outcome of an optimization trap that ensnares even the most successful organizations.

Following the example of natural ecosystems, businesses follow a similar adaptive cycle. Scrappy entrepreneurs race to establish themselves in new market territories, founders shape partnerships, bring on talent, and create enduring customer relationships. As companies mature, they enter the production phase, with efficiency and productivity soaring as processes become refined and systematic. The organization builds deeper expertise, standardizes workflows, and develops sophisticated, revenue-boosting operations. Everything appears in control and thriving.

But then subtle, dangerous changes emerge. Over time, everyone gets used to “the way it’s always been done,” and falls into entrenched habits. Leadership’s focus gradually shifts away from learning and adapting to optimizing the status quo. Ultimately, success’s desire for predictability begets rigidity.

The Optimization Trap

The irony is that the same networks, processes, and cultural mindsets that once ensured success are now roadblocks, stifling the renewal and innovation necessary for long-term survival. What made the company great transforms into vulnerabilities. Instead of experimenting with new approaches, companies instinctively defend their legacy products and established processes, driving out the type of “destructive” change that seeds renewal. Why struggle with the risky, uncertain work of building something new while “the old way” is still profitable?

But when that system is significantly disturbed—by disruptive technology, regulatory shifts, or a shaky economy—the collapse can be swift and devastating. Organizations overly optimized for specific conditions lack the flexibility to adapt when those conditions change. This optimization trap has claimed victims such as Kodak and Toys “R” Us, each following a remarkably similar pattern of initial success, operational perfection, and eventual obsolescence.

Breaking Free Through Mini-Cycles

Companies that thrive over decades understand a crucial principle of fostering organizational mini-cycles long before calamity forces change upon them. Amazon creates autonomous teams that operate like internal startups. Google’s “20% time” policy encourages employees to work on experimental projects outside their main responsibilities. 3M has historically spun off new divisions from promising research initiatives, maintaining entrepreneurial energy within a large corporation.

These practices mimic nature’s approach to resilience. Rather than trying to maintain perfect stability, successful organizations allow for controlled failure and planned renewal. Small, nimble teams, product spin-offs, and dynamic partnerships inspire fresh thinking and challenge established ways of doing business. Some initiatives will fail, but these failures become profitable investments in organizational growth and learning.

Progressive companies also build diverse portfolios of initiatives at different lifecycle stages that ensure the entire organization is never dependent on a single approach or vulnerable to the same disruption.

The Art of Perpetual Renewal

The most enduring businesses don’t simply weather change; they create it. Staying competitive requires constant evolution. Perpetual optimization is a dead end. While their competitors focus on perfecting yesterday’s winning formula, they dive headlong into new opportunities. The Netflix example is an ideal illustration. Rather than protecting their profitable DVD operations, they embraced the discomfort of competing against themselves. Even as their DVD-by-mail service overtook Blockbuster, they simultaneously developed streaming capabilities that would eventually cannibalize their own business model (and go on to forge the world’s largest on-demand media service).

In business, the temptation is always to lock in the current winning formula. But the most resilient companies preemptively build for change. In a world of accelerating change, the biggest risk isn’t failure to optimize; it’s failure to adapt.

In our next post, we’ll turn from strategy to everyday life and explore how the adaptive cycle can help you navigate layoffs, relocations, and relationship shifts—not as crises to endure, but as openings to redesign your identity, renew your networks, and come back stronger.

When I sat down with management and asked what they’d done with the analysis, there was a pause.

They had prioritized sites for climate action based on which ones generated the most revenue.

They didn’t need climate projections to tell them that. They already knew which sites mattered most to the business. The analysis had produced a large volume of impressive-looking output that wasn’t actionable, so management fell back on what they already understood.

What happened here has a name in climate adaptation science: outcome vulnerability. And it’s one of the most common, and expensive, Planet Simple traps I encounter.

The Assumption Buried in the Analysis

Outcome vulnerability is an approach to climate risk assessment that starts with the climate model and works outward. The assumption is that projections of future physical conditions will automatically reveal what needs to be done. Get the science right, and the management decisions will follow.

It’s an appealing idea. It’s Planet Simple in its structure: establish the facts through rigorous research, and the answer will emerge. The problem is that climate projections are generated at global or continental scales. They model physical variables – temperature, precipitation, sea level – but they don’t model the contextual factors that determine whether those physical changes actually matter for a specific business at a specific location.

The climate model doesn’t know about the local infrastructure. It doesn’t know about the community dependencies, the regulatory context, the operational specifics. Outcome vulnerability assumes it doesn’t need to – that the projection is sufficient to drive the decision. It rarely is.

The Alternative

The approach that actually produces actionable insights for climate change adaptation is called contextual vulnerability. It starts not with the climate model but with the business context: What are the critical functions of this site? What are the existing vulnerabilities in its operating environment? What contextual factors – infrastructure, supply, community, regulation – mediate its exposure to climate change?

The climate model then becomes one input among many, rather than the primary driver of conclusions. The question shifts from “what does the projection say will happen?” to “given what we know about this specific context, how might climate change affect our ability to operate here?”

Interestingly, this is exactly what the management team in my opening story was doing when they prioritized by revenue. They were applying a form of contextual reasoning. They just hadn’t framed it that way – and they hadn’t needed to spend several hundred thousand dollars to get there.

The Illusion of Certainty

The reason that outcome vulnerability is a Planet Simple trap is because of the illusion of certainty that quantitative analysis creates.

A climate scenario analysis that produces a table of projected physical risk scores for each of a company’s sites looks authoritative. The numbers imply precision. But climate projections at the scale of an individual industrial site carry substantial uncertainty, and that uncertainty compounds when you try to translate a physical variable into a business impact.

The false precision doesn’t just waste resources – it can actively mislead decision-making. Management may feel they have understood their climate risk because they have a number for it. They haven’t. They’ve replaced a genuine question with a quantified non-answer.

Contextual vulnerability won’t give you a single risk score. It will give you something more useful: an understanding of where your operations are genuinely exposed, why, and what that means for decisions you’re actually in a position to make.

That’s what a several-hundred-thousand-dollar engagement should deliver.

This is part of the Planet Simple Traps series, exploring the tools and frameworks that look rigorous but quietly reinforce the assumptions holding corporate sustainability back. Based on Leaving Planet Simple by Dr. Alex Gold.

That definition of recovery makes sense for a knee. It makes much less sense for an organization navigating a destabilizing world.

But it’s the definition of resilience that most businesses are working with. And it’s a Planet Simple trap.

Two Kinds of Resilience

The concept of resilience entered scientific literature through the work of ecologist C.S. Holling over fifty years ago. Since then, it’s split in two directions that look superficially similar but lead to fundamentally different outcomes.

The first is engineering resilience: the ability to bounce back after a disruption, with faster return to the original state signaling greater resilience. A bridge that doesn’t sway in a storm. A supply chain that recovers in two weeks instead of six. A system that resists change and returns to its prior configuration.

The second is evolutionary resilience: the capacity not just to absorb disruption but to reorganize, adapt, and emerge differently on the other side – in ways that may be better suited to the changed conditions. This version of resilience doesn’t measure success by how quickly you return to the original state. It measures success by how well you navigate to a viable new one.

These aren’t subtly different. They point in opposite directions. Engineering resilience is about resisting change. Evolutionary resilience is about working with it.

Why Engineering Resilience Dominates (and Why That’s a Problem)

Engineering resilience is intuitive to managers raised on Planet Simple. It aligns with the core objective of maintaining stability and control. It’s also easier to measure. Bounce-back time is quantifiable. The capacity to reorganize into a better configuration is not, at least not in advance.

The problem is that engineering resilience, pursued consistently, produces the same outcome as the ultimate Planet Simple trap: a system that’s increasingly tightly optimized and increasingly brittle. A system that is good at bouncing back from predictable disruptions but fundamentally unprepared for the disruption that changes the landscape entirely.

Resilience scientist Lance Gunderson put it directly: engineering resilience “relies on an implicit assumption of global stability” – i.e. the idea that there is one equilibrium to return to. Evolutionary resilience acknowledges that the disruption might mean the old equilibrium no longer exists.

What This Looks Like in Practice

A company whose supply chain was disrupted by COVID-19 and responded by building in more inventory buffer and adding backup suppliers has improved its engineering resilience. It’s better positioned to absorb the same disruption next time.

A company that used the same disruption to fundamentally question its sourcing model - exploring regional supply chains, reassessing which inputs it actually needs, reconsidering the geography of its operations – is working toward evolutionary resilience. It’s not just absorbing the disruption; it’s using it as information about how the world has changed.

Most companies did the former. It was faster, more measurable, and didn’t require questioning any foundational assumptions. That’s exactly the pattern to watch for.

The Question That Changes Everything

Next time you hear a company describe itself as “resilient,” it’s worth asking: resilient to what, and in which direction?

A knee that bounces back quickly is a success. An organization that bounces back to the same configuration that made it vulnerable is an accident waiting to happen again.

This post is part of the Planet Simple Traps series, based on the book Leaving Planet Simple by Dr. Alex Gold. More at dralexgold.com.

This scenario has played out repeatedly. The 2020 California wildfires burned over 4.2 million acres—a direct result of decades of fire suppression. In the same year, Colorado’s Cameron Peak Fire burned over 200,000 acres.

Attempts to control natural water flows have led to similar results. In the Everglades, engineered flood control has disrupted natural water cycles, making both droughts and floods more extreme when they inevitably occur.

These disasters are not accidents of nature. Instead, they are predictable consequences of trying to enforce stability on ever-changing systems—the ultimate trap of Planet Simple.

The Ultimate Trap of Planet Simple

Nature operates on the adaptive cycle—a four-phase process governing everything from forests to wetlands. After disturbance, new growth colonizes cleared space. As systems develop, connections multiply and strengthen. The forest becomes denser, root systems interweave, wildlife communities establish complex networks. Everything becomes increasingly interconnected and productive, reaching peak biomass and maximum output.

This is exactly when natural resource managers—forest rangers, water authorities, and so on—try to hit the pause button. Faced with productive ecosystems that seem stable, we convince ourselves we can maintain this “sweet spot” indefinitely, suppressing fires, diverting waterways, and overengineering crops.

But nature isn’t static. Seeking constant productivity through stability almost always backfires, leaving us victims of our own misunderstandings. Let’s consider some very real examples:

• Fire Suppression’s Dangerous Game: Every year without fire, forests accumulate more fuel. What would naturally be cleared by regular, low-intensity fires instead builds into massive combustible reservoirs. When fire inevitably comes, it explodes into raging crown fires, completely destroying the very forests we tried to protect.

• Flood Control’s False Promise: Engineers install levees and drainage systems to maintain “ideal” (for us) water levels, but natural flood cycles redistribute nutrients, clear accumulated matter, and reset ecological communities. When we eliminate small, regular floods, we set the stage for catastrophic ones that overwhelm our controls and cause far more damage than natural flooding ever would.

• Agricultural Rigidity: Modern agriculture pursues production with monocultures, standardized inputs, and maximum yields. But genetic uniformity creates vulnerability. When new pests, diseases, or climate conditions arise, these rigid systems lack diversity to adapt, resulting in sudden, system-wide collapse.

The more energy we invest in maintaining the illusion stability, the more devastating the fated change becomes. The stability is a trap—the ultimate trap of Planet Simple.

Avoiding the Stability Trap

The most successful resource managers embrace the adaptive cycle rather than fight it, knowing that true resilience comes from managing change intelligently, not preventing it.

Instead of suppressing all fires, for example, forest managers conduct regular prescribed burns during safe conditions to clear accumulated fuel before it becomes dangerous (inspired by the “mosaic burning” of indigenous peoples worldwide). Progressive managers go on to create patches of multi-aged forests—some in early growth, others mature, still others recently disturbed and renewing—to ensure the entire system is never vulnerable to the same threats simultaneously.

In the fields, regenerative farmers avoid uniform monocultures, instead relying on crop rotation, cover cropping, and diversified plantings to reduce vulnerability and maintain long-term productivity.

The lesson is simple: systems that survive and thrive don’t fear change. Instead, change is renewal.

Embracing the Inevitable

Trying to keep nature in stagnant “optimization” isn’t just futile, it’s dangerous. The true art of management lies in the inner workings of the adaptive cycle’s energy. Forests that burn regularly with low-intensity fires remain healthy for centuries. Seasonally flooded wetlands maintain ecological functions through drought and deluge. Agricultural systems incorporating diversity weather unexpected challenges while maintaining productivity over generations.

Nature has been perfecting this dance for millions of years. It’s time we learned the steps.

In my next post, we’ll explore how this pattern plays out in the business world, where the cost of rigidity can be measured in dollars, jobs, and entire industries that vanish when they fail to embrace the adaptive cycle.

That structure — a parallel initiative designed to offset the impacts of the core operation without touching the core operation — is exactly how Corporate Social Responsibility works.

And it’s one of the most consequential Planet Simple traps in business today.

The Name Says It All

CSR as a movement emerged in the 1970s. The name itself reveals the assumption embedded in it: that business is, by default, irresponsible — and needs a dedicated strategy to balance that out. For the era, there was a certain logic to this. Industrial pollution was visible and localized. “Bad actors” could be identified. The solution was cleanup: add a responsibility layer to offset the damage.

The problem is that this cleanup model never left. It calcified into orthodoxy. Today, decades later, most large companies still treat sustainability as a separate function — staffed by a sustainability team, reported through a CSR or ESG report, funded as a cost center rather than integrated as a strategic imperative.

This is the CSR trap. Not because it’s dishonest, but because it’s structurally incapable of doing what it claims to do.

Philanthropy Is Not Sustainability

A lot of what gets called CSR is philanthropy: donations to nonprofits, employee volunteering programs, community partnerships. These activities are often genuinely good. The PR value is high. The emotional resonance is real. But conflating them with sustainability is where things go wrong.

Philanthropy addresses symptoms. Sustainability requires engaging with root causes — how the business model itself interacts with social and environmental systems. When a company funds a tree-planting initiative while expanding a supply chain with deforestation risk baked into it, it isn’t doing sustainability. It’s doing optics. And at some level, most sophisticated observers know the difference.

The deeper problem is what CSR does to the organization’s self-perception. If you have a CSR program, you’ve done sustainability. The work is complete. There’s no impetus to ask harder questions — about supply chain design, capital allocation, how the core value proposition interacts with the environment and communities it depends on.

CSR, in this sense, is an inoculation against transformation.

“Sustainability is in our DNA”

Something I hear frequently: ‘Sustainability is in our DNA.’ It may well be in the mission statement too. What it usually isn’t in is the core budget, the capital allocation framework, or the performance metrics of anyone who runs a business unit.

This is the Planet Simple trap in action.

The premise it rests on — that business and sustainability are distinct domains that occasionally intersect — is exactly the assumption resilience science has disproved. Businesses do not operate outside natural and social systems. They operate within them. The environmental and social context isn’t a nice-to-have; it’s the substrate on which value creation depends.

A company that runs a CSR program while treating sustainability as external to its core business model hasn’t transformed. It has installed a pressure valve. When conditions change — regulatory, physical, or reputational — that pressure valve won’t hold.

What Getting Out of This Trap Looks Like

The companies making the most genuine progress aren’t the ones with the glossiest achievements. They’re the ones that have started asking: How does our business model depend on environmental and social factors to create value? And what does it mean to manage those dependencies deliberately?

That question sounds simple. It requires a completely different mindset to answer — one that treats sustainability not as a function to be managed separately, but as the context in which all functions operate.

You can’t offset your way out of it. You have to change the operation.

I ask them to imagine a planet designed for human convenience. A planet where nature is stable and predictable. Where cause leads reliably to effect – every time. Where complex problems can always be broken into discrete parts, and then built back up to be solved. Where the planet is a setting for economic activity (the backdrop, stage, prop) rather than the system that makes it possible.

That planet doesn’t exist. But we think we’re living in it.

I call it Planet Simple.

The name comes from an acronym that captures its defining assumptions: Stable, Individualistic, Measurable, Predictable, Linear, and at Equilibrium. A set of beliefs about the world that trace back to the Scientific Revolution and still quietly govern how most organizations think about risk, strategy, and sustainability today. The assumptions are deeply embedded. Most of the time, they’re invisible. That’s precisely what makes them dangerous.

Here’s the core problem: our businesses, our institutions, our entire model of economic life were constructed on a worldview that contemporary science has since disproved. Sustainability was meant to be the correction. Instead, it got built into the existing structure – reinforcing the foundation rather than replacing it.

The result is what I call Planet Simple traps — ways that corporate sustainability gets distorted by the very mindset it’s supposed to challenge.

A Planet Simple trap doesn’t look like failure from the outside. It looks like progress. A company publishes a CSR report with all its good deeds. A risk team completes a climate scenario analysis. A company surveys its stakeholders to identify material issues. All the boxes are checked. None of the underlying assumptions have changed.

That’s the trap.

The traps aren’t signs of bad faith. Most of the organizations falling into them are trying to do the right thing. The problem is that they’re trying to do the right thing using the wrong mental model — one that was never designed to handle systemic, nonlinear, deeply uncertain global challenges. You can’t manage a complex adaptive system with tools built for a world of equilibrium and control.

I know this pattern well because I’ve lived it. I’ve been embedded in public and private sector organizations tasked with integrating sustainability into strategic planning. Every time, the barriers aren’t a lack of knowledge or willingness — they were structural. They were Planet Simple at work.

My experience was the basis for Leaving Planet Simple — and now this series. All in the hope of laying the groundwork for a better future.

Over the coming posts, I’ll walk through ten specific Planet Simple traps, drawn from the research and case studies in Leaving Planet Simple. Each one distorts a genuine method — resilience, scenario analysis, materiality, reporting — by fitting it into an existing Planet Simple framework rather than allowing the method to challenge that framework. Each one allows practitioners to claim they are doing sustainability while preserving the assumptions that gave rise to today’s global challenges in the first place.

The ten traps (in no particular order) are:

1. Corporate Social Responsibility: how the cleanup strategy became the strategy

2. Engineering Resilience: the definition of resilience that makes your organization more fragile

3. Outcome Vulnerability: why more climate data doesn’t mean better decisions

4. Scenario Analysis or Sensitivity Analysis: the difference that most companies miss

5. Adaptive Management or Incrementalism: when learning from experience is just muddling through

6. The Double Materiality Matrix: what’s wrong with putting sustainability issues in a box

7. The Most Likely Scenario: why identifying it sets your strategy up to fail

8. Aggregating Climate Risk: the averaging that produces the wrong answer by design

9. Substituting Means for Ends: when the analysis becomes the destination

10. The Ultimate Trap of Planet Simple: why optimizing a system is the surest way to break it

You don’t need to be a sustainability professional to follow this series. If you’re a business leader who senses that something is off about the current sustainability conversation, i.e. that there’s a lot of activity but not enough transformation, this series is for you.

The alternative way of seeing this has always been there. Let’s find it.

This post is part of the Planet Simple Traps series, based on the book Leaving Planet Simple by Dr. Alex Gold.

The only constant in life is change, and we can attempt to defy it, reshape it to something more to our liking, or acclimatize and roll with those changes. From the field of resilience science was born the adaptive cycle, a model that reflects real-world behavior and circumstances of human-natural systems.

Unlike linear models of change that see progress as a straight line, or cyclical models that simply move back and forth, the adaptive cycle reveals change as a dynamic four-phase process operating simultaneously across multiple periods of our existence—growth, production, destruction, and renewal. At its core, the cycle shows us that destruction isn’t the opposite of growth but rather a necessary precursor to renewal. This reframes our relationship with change from something to resist to something to embrace as a natural part of life’s rhythm.

The adaptive cycle in motion

Consider an ecosystem example where a scattering of white pine saplings colonize recently disturbed or otherwise vacant space. The trees grow, building up natural capital in connectedness with other resources. The same dynamic occurs in business environments, families, and social networks where the more you connect and build, the more productive you can be.

At some point, a system will reach a climax state of high connectedness and high potential. Maybe it’s a forest with substantial biodiversity, a fishery yielding major catches, or a business operating at maximum efficiency. Because of the connections and experience built up in the system, its potential is huge. As will be explained later, the trap is trying to keep a system in this highly productive state indefinitely.

But then something happens that throws a wrench into the works, such as a forest fire, from which existing capital is released.

While dramatic and potentially devastating, the destruction phase naturally breaks up a system, sowing the seeds of potential productivity toward a place of renewal (species, participants, ideas) that sets the stage for the next growth cycle. Here’s a related example: I recently moved from New York City to Washington, D.C., breaking up connections in one place to see opportunity to join new social clubs or experience things I didn’t have before, embedded in an existing system. And that is a key element people sometimes forget when going through change—when one door closes, three others open. However, you can only join so many new social clubs or business groups or adventure groups. At some point you need to commit to something fulfilling in order to launch your next cycle of connectedness and potential.

Beware the trap

Keep in mind that this cycle is not without its snags. Imagine a business that has invested significant time and capital into building connections and productivity to a high level. Most company leaders wish to maintain that productive, largely lucrative, pace, but due to the constancy of change, it’s impossible to stay the same indefinitely. The more you try to keep something the same, the harder the fall when crisis comes. Every business strives to innovate; however, innovation requires change. You can have maximum production or reserve some energy to innovate, but not both. Creative destruction lets you create mini-cycles to sustain your business at scale.

Therein lies a key lesson of resilience science: stability is an illusion. The production phase and its outputs are expected to remain stable, but imposing a static goal on a dynamic system is bound to fail.

Seeing the adaptive cycle in everyday life

What does it look like to apply this cycle in practical ways—at work, at home, and in our communities? How can we use resilience science to harness creative destruction, foster renewal, and build stronger, more adaptive lives and enterprises?

Stay tuned for the upcoming series, where we will shift our discussion of the adaptive adaptive cycle from metaphor to reality – with everyday examples from business to the personal.

There’s a persistent narrative in corporate sustainability circles that we’re just one more dataset away from solving our nature crisis.

The storyline goes something like this: if only we had better measurements, more comprehensive biodiversity assessments, and more granular ecosystem data, businesses could finally address their nature-related impacts and dependencies. This belief has spawned countless initiatives aimed at closing these nature “data gaps.”

The nature challenge facing business isn’t a lack of data. It’s our fundamental miscalculation of what nature is and how it works. When we talk about “filling nature data gaps,” we’re operating under a Planet Simple mindset: the misguided belief that nature is something external to business that can be fully measured, quantified, and ultimately controlled.

A Relationship of Illusion

Consider that scientists still cannot agree on how many species exist on Earth. Current estimates range from 2 million to over 1 trillion species, with a “best guess” somewhere between 8 and 20 million. If the scientific community—armed with sophisticated tools and centuries of taxonomic research—cannot reach a consensus on this seemingly straightforward question, is it realistic to believe corporations can accurately assess their full impact on nature?

This pervasive uncertainty is built on the fact that nature is not a static entity waiting to be cataloged; it’s a complex, dynamic web of relationships that defies our attempts at complete understanding. Whenever we think we’ve mapped out an ecosystem, we discover new species, interactions, and complexities that challenge our previous assumptions.

Yet corporate sustainability initiatives often proceed as if nature were a puzzle missing just a few pieces. We create sophisticated reporting frameworks, develop “comprehensive” nature assessment tools, and build elaborate databases—all operating under the illusion that with enough data points, we can fully grasp and manage our relationship with the natural world.

We do all this – and yet the multitude of metrics has often led to confusion and inaction rather than meaningful change

The Actual Nature Data Gap

Instead of looking outward, we need to look inward.

Businesses need to shift their focus from attempting to quantify nature through complex metrics to understanding how their current business models are fundamentally intertwined with natural systems. Are we acknowledging how our business depends on natural resources, and how our activities impact natural systems? Or are we too busy waiting for more data about the “state of nature”?

When we look inward, we focus on unpacking the reality that has been hiding in plain sight. Leading guidance helps businesses make the conceptual shift by defining impact drivers that businesses can focus on. Some of these impact drivers are:

• Land use change – such as converting forests to agriculture or clearing land for property development

• Resource use – including water use, minerals, and direct use of natural resources such as forestry

• Pollution – air, water, and soil pollution that affects natural habitat

Most businesses know whether they’re clearing land, increasing resource use, or releasing pollutants. But few have connected these activities to “nature.”

When businesses look inward and use impact drivers to define the business-nature connection, they can immediately identify a few ways to manage nature impacts by addressing what they’re already doing. This is far more straightforward than hoping “nature data” will provide the answer.

Nature Isn’t Just ‘Out There’

The next frontier in corporate sustainability isn’t more nature data—it’s nature intelligence. This means developing the capacity to think systemically, act with humility, and make decisions that acknowledge both the known and unknown dimensions of our relationship with the natural world. Only then can we move beyond the Planet Simple mindset and toward truly sustainable business practices that honor the complexity and mystery of the natural systems upon which we depend.

Business has always been part of nature and depends heavily on its resources. We don’t need more data about nature “out there”—we can now start understanding our connections to nature. With this intelligence, business has what it needs to thrive as our planet evolves.

I’ll be the first to say I hope my headline is wrong. But let’s explore the facts and then do a thought experiment.

As California grapples with yet another round of devastating wildfires, the Golden State finds itself at a critical juncture. The recent infernos that ravaged Los Angeles County, destroying thousands of homes and displacing over 100,000 residents, are just the latest in a string of climate-fueled disasters that have battered the state in recent years.

The intensifying cycle of wildfires, droughts, and floods has not only taken a toll on Californians’ lives and property but has also sent shockwaves through the insurance industry. Major insurers, faced with mounting losses and an increasingly unpredictable risk landscape, have begun a mass exodus from the California market. State Farm, Allstate, and others have either severely limited their coverage or pulled out entirely, leaving millions of homeowners in limbo.

In response, California has been forced to expand its FAIR Plan, the state-run insurer of last resort. However, this stopgap measure is proving increasingly inadequate as climate change amplifies the frequency and severity of natural disasters. The FAIR Plan, originally designed to provide basic fire insurance for those who couldn’t obtain it elsewhere, now finds itself overwhelmed by the sheer scale of the crisis.

Let’s Turn to What Might Happen

As the state struggles to keep up with the spiraling costs of disaster recovery, it’s becoming painfully clear that California’s resources are stretched to the breaking point. The billions of dollars required to rebuild communities, restore infrastructure, and mitigate future risks are simply beyond the state’s capacity, especially as it faces other pressing challenges such as housing affordability and water scarcity.

In this dire situation, Californians have turned to the federal government for assistance, as they have done in past crises. However, they now face an unprecedented obstacle: a hostile White House led by President Donald Trump.

During the first Trump presidency, aides had to pressure him to release federal disaster aid to California after the 2018 wildfires after he initially refused. Now it’s Trump’s second term. California still isn’t a fan, and an emboldened Trump doesn’t give California a cent of federal aid. Fueled by his long-standing feud with California’s leadership, particularly Governor Gavin Newsom, Trump defends the decision on social media with inaccuracies about California’s policies and FEMA funding.

This political brinkmanship strikes a nerve with Californians, who are acutely aware of their outsized contributions to federal coffers. As the world’s fifth-largest economy, California consistently pays more in federal taxes than it receives in federal spending. In 2021 alone, the state contributed a staggering $472 billion in federal tax revenue. Now, despite their significant financial input, the federal government has abandoned Californians in their time of need. Californians feel betrayed and exploited.

As frustration mounts, a once-fringe idea regains traction: secession. The “Calexit“ movement, which has ebbed and flowed over the years, experiences a strong resurgence. Buoyed by the urgency of crisis and the emotion of injustice, a coalition of grassroots organizations and state legislators successfully places a secession initiative on the ballot.

The proposal would redirect California’s federal tax payments to a state-managed fund, effectively cutting off the flow of money to Washington while retaining resources for disaster response and recovery. Californians design a future where their tax dollars directly benefit their state rather than being held hostage to federal whims.

Retaining a-half-a-trillion dollars a year to rebuild the state is an easy sell. The Golden State is now the Golden Nation.

California is the first domino to fall, as a polarized nation moves to take the economic reality of climate change into its own hands.

Let’s Finish With Some Facts

Climate change is here.

Climate change is a financial issue.

Addressing climate change is an economic imperative for America.