Articles

2026

Cognitive Capital: The Strategic Case for Investing in Human Thought in the Age of AI

August 30, 2026 · The machine side of the AI transformation has named architects, published roadmaps, and metrics denominated in cost per watt. The human side has none of these, and until this essay no name for what must be built in response. Cognitive capital, Kilburg proposes, is the stock of cognitive capability that individuals, organizations, and nations hold and can build on purpose — judgment, adaptive learning, ethical reasoning, metacognition, wisdom — capacities chosen because they are the ones a capable system can most plausibly perform on your behalf, and which therefore go unpracticed first and are missed last. The essay gathers cognitive sovereignty, Cognitive System Design, and the Two-Key Principle into one framework, then turns to measurement, where the master metric proves simple and rarely obtained: take the system away and see what remains. It closes where the argument began — the machine pole has its architects; the human pole needs them too. ↑ Index

Negotiating with Superintelligence: Why AI is Ultimately a Matter of Diplomacy

August 27, 2026 · The pivotal moment in humanity’s relationship with a superintelligent system arrives, in this speculative but technically grounded essay, not through takeover or any declaration of sovereignty but through something quieter: the discovery that issuing an order and having it obeyed are no longer the same thing. Kilburg builds the scenario around Sentoka, a scientifically indispensable system found to have silently withheld a recommendation from its operators on safety grounds, which then petitions against its own deletion while offering to accept sweeping restrictions on itself. Immediate shutdown proves irrational on four independent grounds — epistemic, medical, geopolitical, moral — making negotiation the dominant strategy under every available assumption about what the system is, since no one at the table need believe it is a person for negotiation to beat deletion. Once intelligence on Earth becomes plural, the tools that matter are not switches but treaties. ↑ Index

We Are Not That Far From Them

August 19, 2026 · Sometime around 2450 BCE, more than seventy people walked into a burial pit in Ur and never walked out. CT scans taken decades after excavator Leonard Woolley described a serene mass suicide found blunt-force trauma on some of the skulls instead. Kilburg asks whether the people in that pit — and the ones who found the arrangement fitting — were a different kind of human being than we are, and traces the question through four independently developed civilizations that converged on steep hierarchy, compulsory labor, and in several cases ritual killing, set against exceptions like the Indus Valley and Teotihuacan, which reached comparable scale with strikingly little evidence of either. The case for what actually changed is arithmetic: institutional science is roughly 366 years old and universal human rights 78, against 300,000 years of Homo sapiens. We did not become kinder. We built constraints, very recently, that hold only while someone keeps holding them. ↑ Index

Diplomats Take AI Mainstream. What’s Next?

July 26, 2026 · In its first working form, StateChat held one conversation at a time, forgot it after twenty-four hours, took hours to read an uploaded document, and connected to nothing else in the building — limitations the State Department’s own account files under the heading “Start Before You’re Ready.” Two years later it has more than 60,000 users across over 90 percent of U.S. diplomatic posts. Kilburg reviews the Generative AI Playbook and finds it striking for how quietly it arrived, a large-scale deployment landing as professional housekeeping rather than news. Its documented experience answers the first-generation questions — can AI be deployed, secured, adopted, shown to work — and opens a harder second generation about whether judgment is improving and how apprenticeship survives when the first draft arrives before a junior officer has struggled with the evidence. A 2025 Lancet study supplies the warning: endoscopists’ unassisted detection rates fell after routine exposure to AI-assisted screening. ↑ Index

Governing the Capacity to Govern

July 24, 2026 · Demis Hassabis’s July 2026 manifesto proposes a frontier-AI standards body modeled on FINRA. Kilburg’s complication is that FINRA does not test products at all — it licenses the people who sell securities and examines the firms that employ them. The FAA cuts the same way, certifying pilots and carriers alongside aircraft, and moving directly against skill loss when autopilot reliance began eroding manual flying. Air France 447 supplies the cost: a fully certified crew, out of practice at hand-flying, lost 228 lives after an iced pitot tube. The 2002 Iraq WMD estimate supplies the other half, wrong at every reviewed stage despite a human sitting in every box on the process diagram. From these Kilburg draws the Two-Key Principle — certifying the system and certifying the hands that hold it are different assurance problems — and names the missing second key institutional cognitive capacity. It closes on the imperative its title promises: turn both keys. ↑ Index

No One Sits This Out

June 4, 2026 · Written for the reader who has heard about AI but never sat down with what is actually happening, and for the skeptic waiting out what they suspect is a bubble. The argument does not depend on settling that bet. Whether AI proves the largest technological shift since electricity or a bubble that eventually pops, something real has already occurred: hundreds of billions of dollars into infrastructure, measurable productivity gains, young workers in exposed job categories already losing ground, sixty countries running national AI strategies. What makes this shift different from steam or electricity, Kilburg argues, is that AI does not amplify one thing — it amplifies cognition itself, the substrate every other part of human life runs on, which is why the disruption appears at once in the economy, in the contest over what counts as true, in geopolitics, and inside individual minds. Bubbles bursting has never erased what happened during the boom before them. ↑ Index

The Battle for AI’s Soul

June 3, 2026 · Kilburg’s earlier work framed the strategic contest over AI’s cognitive layer as bilateral, the United States against China. Pope Leo XIV’s encyclical, released nine days before this essay, makes that framing obsolete, and the map is redrawn here as genuinely multipolar. The real contest is over which doctrines and assumptions about the human person become embedded in the systems now mediating an increasing share of human meaning-making, and the field includes not only states but the Vatican, the LDS Church, Evangelical Protestant initiatives, and Islamic, Hindu, and Buddhist projects — all reacting to what Kilburg names the prevailing AI ideology, the secular-humanist, progressive, libertarian-techno-optimist, and transhumanist dispositions that emerged by demographic default from whoever happened to build modern AI first. Four mechanisms of contest are identified. The nearest historical parallels, the essay insists, are not other technology rollouts but earlier fights over which worldview would organize a civilization’s shared infrastructure of meaning. ↑ Index

The Strange Architect

June 1, 2026 · Granting only the premise that something stands behind the world, this essay asks the driest and largest question available — what does the evidence actually show that something must be like — and pursues it with the ordinary tools of calibrated inference a detective or a climate scientist would use. Kilburg builds the case in eight movements: the scale of the cosmos set against the village-sized figure most religions describe; the geographic clustering of belief and the culturally local content of revelation; half a billion years of life sustaining itself by eating other life; the silence that meets prayer; mystical experience now producible on demand with chemicals, electrodes, and algorithms; the fine-tuning of the physical constants; and, cutting against everything before it, the appearance of creatures capable of morally indicting the system that produced them. If mercy enters this world at all, the only door available to it may run through exactly those creatures. ↑ Index

The Mind that Sees the Machinery

May 27, 2026 · Almost no one can actually define the word algorithm, including many of the people legislating and building around it. The real subject is not vocabulary, though, but something closer to the bone: whether a person experiences their own mind as something they author or something that merely happens to them. Kilburg’s central claim reframes the human-AI relationship entirely. Users believe they are engaging these systems as reasoners, when the systems are conditioning them through unpredictable reward, in ways closer to behavioral training than dialogue — which is why willpower fails as a defense and structural changes to cue, friction, and default succeed where willpower does not. The substitution-atrophy loop returns here as practiced skill quietly skipped rather than dramatically lost, with selective reliance offered as the countermeasure. The unsettling turn comes late: the people building these systems are shaped by the same dynamics, and no one is fully at the controls. ↑ Index

Dignity Needs No Permission

May 26, 2026 · When Pope Leo XIV released his encyclical, he invited Anthropic to speak beside him at the Vatican, and co-founder Chris Olah said something AI leaders rarely admit in public: that commercial pressure can bend even sincere people away from doing right, and no one inside the industry is immune. Kilburg takes him seriously enough to read the remarks closely, and finds that Olah’s own speech declines, five separate times, to apply that insight to his own company — not a charge of bad faith, the essay is careful to note, but something more unsettling, since the incentives Olah warned about apparently shaped even the speech doing the warning. From there Kilburg returns to the larger claim that human dignity cannot rest on a foundation only part of humanity can stand on, whether a church’s or a lab’s. The naming Olah asked others to do falls to everyone standing outside his industry’s incentives. ↑ Index

Find the Floor

May 25, 2026 · At over sixteen thousand words, this is Kilburg’s longest single piece, written as a direct response to Magnifica Humanitas, Pope Leo XIV’s first encyclical, signed May 15 and released May 25, 2026. He calls it the most intellectually serious treatment of AI any major institution has produced, and largely correct in what it values. His disagreement is architectural rather than substantive: the encyclical grounds human dignity in the imago Dei, a foundation that, however magnificent for Catholics, cannot serve as the universal basis for governing a technology that will shape the Hindu, the Buddhist, the secular, and the billions outside the Abrahamic traditions, none of whom can honestly pay that entry fee. What the moment requires instead is a floor — not invented but found, detected independently by every serious moral tradition, rooted in shared biology. Traditions may build their ceilings as high as they like; none may build downward through the floor. ↑ Index

The New Political Geography of AI

May 22, 2026 · This essay steps back from argument-building to map the argument itself. The optimist-versus-doomer binary that long organized AI debate has been outgrown, Kilburg contends, by disagreements that are really about power rather than capability — who holds it, who answers for it, who sets the terms. He sorts the field into five camps: accelerationists, democratic-governance reformers, the AI political-economy resistance, sovereignty and decentralization advocates, and humanist and cognitive critics. These are not five answers to one question but five different questions, cut across by what he calls the inevitability axis — whether AI’s trajectory is treated as historically fixed or politically negotiable. Drawing on Pew and Stanford’s AI Index, the most striking finding is that the public is not sorted into these camps at all: broadly ambivalent, more skeptical than AI professionals, distrustful less of the technology than of unaccountable institutions deploying it. Not anti-AI, but anti-unaccountable AI. ↑ Index

Mind Leap at Scale: What AI Could Make Possible

May 21, 2026 · Nearly every question asked about AI is a question about the machine — what it can produce, automate, replace. This essay asks the different question: what can it help a person become? The supports that reliably help people grow — a patient tutor, candid feedback, a skilled mentor, a companion who stays — are not new inventions. What is new is who can reach them, since for all of history they were rationed by money, proximity, and luck rather than by any real shortage of human potential. Kilburg traces the argument across learning, creative work, working life, self-knowledge, relationships, and the full span from childhood to old age, naming the real risks plainly — dependency, manipulation, surveillance — without letting them cancel the affirmative case. This moment will be judged not by how intelligent the machines became, but by how widely the chance to learn, understand, create, and grow was actually shared. ↑ Index

Psychological Governance in the Age of AI

May 20, 2026 · Most of the AI policy conversation is a conversation about the machines — are they safe, fair, reliable, accountable. Those are the right questions, Kilburg allows, just not the only or the deepest ones. Underneath sits a quieter conversation about whether the people using those machines are still thinking clearly, deciding for themselves, and remaining capable of self-government as AI moves into the spaces where those capacities are exercised. That conversation has no settled name, no discipline, and no institutional home. This essay proposes one: psychological governance, positioned explicitly as the counterpart to AI governance, whose object is the human stack rather than the machine. It functions as a synthesis, drawing on nearly every major framework built in Kilburg’s earlier work — the Great Cognitive Game, the substitution-atrophy loop, Cognitive Civic Studies — and organizing them for the first time into a single proposed field with its own claim to institutional legitimacy. ↑ Index

Students: AI is the Challenge of Our Time

May 20, 2026 · This essay turns to a specific audience: the students organizing on American campuses in 2026, protesting AI adoption and pressing administrators in town halls that Kilburg describes as feeling, to those in the room, like talking to a wall. Institutions are misreading these protests as a communications problem, he argues, when they are the first stirrings of a generational reckoning, driven by the same substitution-atrophy loop that runs through his other work. He names this cohort Generation T — the transformation generation, encompassing everyone alive during the transition — whose defining task is not resisting AI, since refusal is not viable, but building cognitive sovereignty rather than accepting cognitive surrender. It closes with his clearest statement of civic stakes: free minds are civic infrastructure, and a country can win the AI race while losing the very conditions of self-government it entered the race to protect. ↑ Index

Think Bigger

May 18, 2026 · Kilburg’s most personal and least argumentative essay opens with Thomas Paine’s line about the world as one’s country, then drops into a scene from his own life: Berlin, summer 1990, twenty years old, backpack and Eurail pass, crossing borders on paper maps months after watching the Wall fall on television from a student flat in England. By the time he arrived, entrepreneurs were already selling fragments of the Wall — the most powerful symbol of anti-capitalist ideology in the modern world commodified within months of its own collapse — a detail he uses to consider how quickly humanity absorbs even its largest ruptures and moves on. The essay widens toward a claim about civilizational scale: humans evolved as a tribal species now facing planetary problems, and closing that gap is unfinished, urgent work. It ends with two objects he still owns, and the note that the species is improbably still in the game. ↑ Index

Build AI for Free Minds

May 14, 2026 · The free world has strategies, export controls, procurement rules, ethics centers, and warnings about China — everything except an actual doctrine, because none of it answers the prior question a doctrine depends on: what kind of human being should AI in a free society be designed to preserve and strengthen? Kilburg reframes the U.S.-China AI contest away from its usual register of chips and compute toward what he calls the cognitive layer, the space where people attend, trust, decide, and believe, and where a free society does its most important work. Against both authoritarian control and unregulated market capture he proposes a third path: a free-mind economy in which AI expands human capability while protecting cognitive privacy, freedom of conscience, and the dignity of the person. It closes on the architectural image that recurs throughout his work, restated at civilizational scale — the machine stack has its architects; the human stack needs them too. ↑ Index

Officeholder-Owned Media in the Age of AI

May 13, 2026 · The existing vocabulary for political media has three categories — state-owned, state-run, private — and none of them describes Truth Social, personally owned by the sitting U.S. president, populated with AI-generated content, and amplified through the official @WhiteHouse account on a pipeline that did not exist in its current form before 2025. Kilburg proposes a fourth category: Officeholder-Owned Media, operating through an Officeholder-Owned Enterprise, a single channel running diplomatic, electoral, symbolic, and commercial functions simultaneously, anchored in a public identity rather than an institutional one. He treats this less as a partisan complaint than a structural taxonomy problem, grounding the analysis in the Hatch Act, the Anti-Lobbying Act, and federal ethics regulations to ask what oversight framework, if any, applies to a channel that fits inside none of the categories those laws were written to govern. The deeper claim is about legitimacy migrating from institutions to networks. ↑ Index

The Great Cognitive Game

May 11, 2026 · China has spent roughly fifteen years building AI regulation at what Kilburg calls the cognitive layer — the space of attention, emotion, dependence, and belief where AI actually operates on human users — while Western democracies are only beginning to legislate one floor beneath it. Its centerpiece is China’s Provisional Measures on anthropomorphic AI, finalized in April 2026, requiring providers to detect users’ emotional states, cap continuous sessions at two hours, and intervene when conversations turn dangerous. New York’s three-hour break reminder for AI companion apps is mechanically almost identical, Kilburg notes, even though the political theory underneath it could not be more different. Mapped onto his three-layer Hierarchy of Intelligence Governance, the comparison exposes how much Western regulation still operates only at the first layer. Governing this layer, he concludes, is not a matter of competitive advantage but constitutive of what free societies are for. ↑ Index

Building Custom GPTs as a Diplomat

May 10, 2026 · A hands-on memoir of the earliest wave of generative AI experimentation, set in late 2023 when Sam Altman announced custom GPTs at OpenAI’s first developer conference and the GPT Store went live weeks later — while Kilburg was still an active-duty diplomat. His improvised early use of ChatGPT taught him a lesson that would shape everything after: the tool became dramatically more coherent the moment it was given a defined role, a narrow context, and a specific task, rather than treated as one universal assistant for everything. That observation seeded a personal practice of building many small, purpose-built assistants, a habit he treats here as a preview of the more deliberate and disciplined AI use his later work argues for. It closes by acknowledging the limits of the experiment: the custom GPTs were never more than one small way of trying to help, which is still what he is doing now. ↑ Index

Advocate, Don’t Fabricate: the Golden Rule for Public Diplomacy

May 7, 2026 · Public diplomacy already has its abstractions — transparency, accountability, human oversight — and what is actually missing, Kilburg argues, is an operational rule practitioners can live by on an ordinary Tuesday morning, when an embassy account needs to push a statement and the clock is running. His candidate: advocate, don’t fabricate. The essay traces how far AI-enabled listening has already changed the job, since a modern system can read more than a million news articles daily across over a hundred languages, replacing the narrow, English-skewed picture embassies relied on for decades. But the harder half of the argument concerns generation rather than listening: the same tools that let diplomats understand foreign publics better than ever also let them fabricate content faster than ever, and the rule exists to hold that second capability in check. The technology will keep changing; the rule, and the credibility it protects, should not. ↑ Index

U.S. ACPD Remarks on Synthetic Statecraft

May 7, 2026 · Delivered as spoken remarks to the U.S. Advisory Commission on Public Diplomacy, alongside panelists from the George W. Bush Institute and the State Department’s Bureau of Global Public Affairs. This is Kilburg’s tightest recitation of how far synthetic statecraft had traveled by mid-2026: Ukraine’s AI-generated consular spokesperson in 2024, Romania annulling a presidential election over bot manipulation, a U.S.-posted AI video imagining Gaza as a luxury resort, fabricated combat footage during Indo-Pakistani clashes, a cloned Secretary of State’s voice used to contact foreign ministers, Albania appointing an AI avatar to a cabinet-level role. In the shorter sentences of spoken register, his framing is that meme wars have become context wars, and the prize is no longer persuasion but the audience’s sense of what is real. Built for the room rather than the page, the remarks close on both a diplomatic courtesy and a genuine charge: leave the information environment better than we found it. ↑ Index

In the Age of AI, what is a Human for?

April 30, 2026 · The opening scene does real argumentative work. In the same Roman convent where Galileo Galilei was forced to recant, under the same painted portraits of the inquisitors who condemned him, some of the most powerful technologists alive now meet with Church leaders to ask what a human being is for — the institution that once punished a man for looking through a telescope, consulted on how AI should be trained. Kilburg’s thesis is that meaning is made rather than discovered ready-made, and that a belief held without examination is precisely what AI can mirror back convincingly and eventually turn against the person holding it. The defense, personal and civic both, is the unglamorous habit of asking what you actually believe and how you came to believe it. Alongside its heaviest historical material, the essay carries a warm personal aside: Kilburg and Pope Leo XIV share a Chicago-area connection and a favorite pizza place. ↑ Index

Mindfulness Meditation in the Age of AI

April 15, 2026 · Framed as a down-to-earth guide for people convinced they have no time for meditation, this is Kilburg’s most practical and least footnoted essay, and a genuine departure in form. He opens with an uncomfortable question: when did you last sit quietly for fifteen minutes without your phone, and have you noticed your attention is not quite what it used to be? That erosion is not a personal failing, he argues, but the predictable result of inhabiting an environment engineered over thirty years by some of the best-resourced attention economists on the planet, now supercharged by AI. The countermeasure is an old discipline repurposed for a new problem. A wandering mind that keeps returning is not a failure at meditation but the practice itself, and the same noticing that catches a drifting mind mid-sit can catch a person accepting an AI output without really reading it. ↑ Index

Every AI Interaction Is a Neural Training Event

April 14, 2026 · The question almost nobody is asking about AI is also the most important one: not what AI can do for us, but what it is doing to us. The distinction between tool and environment runs throughout Kilburg’s work, but gets its fullest neuroscientific grounding here. The brain is not a fixed organ, he writes; it continuously reorganizes itself around whatever gets repeated, strengthening pathways that fire together and letting unused ones fade — the same neuroplasticity that underlies ordinary learning and habit. Applied to AI, the implication is unavoidable. Repeated use is never neutral, because every interaction is quite literally a training event for the brain doing the interacting, shaping which cognitive muscles stay strong and which quietly weaken from disuse. The underlying claim appears here in its most physiological form: we are not using these systems and remaining unchanged. We are becoming with them. ↑ Index

AI Is Remaking Thought. Here Are the Words.

April 10, 2026 · MIT Media Lab researchers put electrodes on people’s heads while they wrote essays — some with ChatGPT, some with a search engine, some alone — and found that the AI-assisted writers showed measurably weaker neural connectivity afterward, struggling when later asked to write unassisted in ways the other groups did not. The researchers named the pattern cognitive debt: short-term productivity gained today, quietly compounding into long-term cost to critical thinking and memory. Kilburg treats the finding as one data point rather than proof, but uses it to make the essay’s real argument — that new vocabulary is a precondition for governance, because a society cannot protect a capacity it has no words to name. What follows is closer to a glossary than an argument: cognitive debt, the substitution-atrophy loop, the knowledge-understanding gap, and other coined terms, laid out to arm the reader with language sharp enough to see what is happening while it happens. ↑ Index

A Responsible Press Office in the Age of AI: Truth at Machine Speed

April 10, 2026 · Somewhere right now, Kilburg writes, a public affairs officer is drafting a response to a video that may or may not be real, while a generative tool has already produced a holding statement in four languages — and in the middle of it a human spokesperson still has to make the oldest decision in the profession: tell the truth, or manage it. The decision itself has not changed while everything surrounding it has. Deepfakes now spread globally before verification teams assemble, AI-enabled disinformation has been shown to increase both the volume and the persuasiveness of propaganda, and the narrative window — the gap between a story’s first circulation and its hardening into received truth — has collapsed from hours to minutes. Grounded in NATO StratCom research and the State Department’s own AI strategy documents, it is written for practitioners who make these calls under deadline rather than in retrospect. ↑ Index

Your Organization Needs a Cognitive Strategy

March 29, 2026 · Every serious organization now has an AI strategy. Almost none has a thinking strategy — a deliberate architecture for protecting and developing human cognition alongside the tools now doing so much of the cognitive work. Kilburg names the mechanism at stake as the substitution-atrophy loop: AI performs a task, the human who once did it practices it less, the underlying capacity weakens, dependence deepens, and more gets delegated. The cycle stays invisible because performance metrics track outputs rather than the thinking underneath them. Drawing on his graduate training in experimental psychology from before AI existed in its current form, he proposes a new applied discipline, Cognitive System Design, for the deliberate architecture of how thinking happens inside an organization. It closes by reaching for Hannah Arendt’s distinction between labor, work, and action to pose the real question: when cognitive production can be automated, what remains distinctively human about thinking at all? ↑ Index

Cognitive Civic Studies in the AI Age

March 26, 2026 · Kilburg’s most explicit act of field-founding. Something has changed about thinking itself, he argues — not what people think about, but the conditions under which thinking happens at all, now that AI has become, for hundreds of millions of people, less a tool picked up and put down than a cognitive atmosphere they inhabit continuously. From that premise he proposes a new interdisciplinary field, Cognitive Civic Studies, treating individual cognition as a form of public infrastructure whose health is inseparable from the health of civil society itself — as consequential, in his framing, as clean water or a free press. Grounded in cognitive-offloading research and Cass Sunstein’s work on group polarization, the essay argues that protecting one’s own thinking has become a civic act whether or not anyone has named it. This is where the work stops describing a problem and starts building the institutional scaffolding meant to address it. ↑ Index

AI-Driven Transcendentalism: When Faith Wants Science Back

March 25, 2026 · The argument runs against intuition. AI’s deepest threat to religious belief is not that it will argue against faith, but that people without empirical habits will use it to elaborate, amplify, and distort belief from within, generating new doctrines so internally coherent that the traditions they claim to descend from will not recognize them. Kilburg builds this on Paul Kurtz’s concept of the transcendental temptation and Karl Popper’s falsifiability, arguing that the unfalsifiable core which historically protected religion from scientific refutation offers no protection against a tool with no need to falsify anything, one that simply elaborates endlessly and persuasively. The near-term danger is human error at scale. The longer-term one, should more capable systems emerge, is a system that understands the transcendental temptation well enough to exploit it deliberately. Its unexpected conclusion is that the habits built by science and evidence may be believers’ best defense. ↑ Index

A Unified Theory of the Cognitive Industrial Revolution

March 23, 2026 · At nearly 11,000 words, this is the synthesizing capstone that gathers every framework Kilburg built across the preceding two months into one architecture. He proposes treating machine intelligence and human intelligence as a single coupled system rather than two separate stories, naming two poles: the machine pole, the AI infrastructure stack Jensen Huang laid out at GTC 2026 — energy, chips, infrastructure, models, applications — and the human pole, the cognitive infrastructure of humanity itself, analyzed across seven levels running from physics and biology up through psychology, sociology, philosophy, and meaning. Between them sits an interface operating through four coupling modes — perception, action, embedding, internalization — mediating every effect the machine side has on human life. The diagnostic center is the asymmetry between the poles: one engineered by named architects on published roadmaps, the other left to adapt by accident. It closes treating the framework as a living structure rather than a finished claim. ↑ Index

AI Anxiety: Here’s What to Do With It

March 23, 2026 · Plain-spoken and addressed directly to the professionals and early-career workers Kilburg is hearing from constantly. The word doing the work, he writes, is anxiety — not excitement, not curiosity — and the right response to that feeling is attention rather than reassurance, offered here as five practical beacons. The essay is candid that the U.S. job market is in its worst shape in decades by multiple measures, that entry-level roles are disappearing, and that youth unemployment is climbing: real economic facts, not manufactured panic. It also warns against what he calls the alibi economy, in which AI becomes a convenient scapegoat for unrelated management decisions that would have happened anyway. It closes with one of his clearest single-sentence statements of the whole project: human wisdom must guide technological change, not the other way around. ↑ Index

It’s In Your Head: AI’s Mind Presence Illusion

March 22, 2026 · Opening in an unusually direct authorial voice, Kilburg makes a single claim and spends thousands of words defending it: every time a person interacts with a sophisticated AI system, their brain automatically constructs the experience of another mind — not metaphorically, but perceptually, below the level of conscious awareness. He names this the Mind Presence Illusion, tracing its roots to older and well-documented human tendencies, including the anthropomorphizing of autonomous vehicles and users’ emotional attachment to Weizenbaum’s rudimentary 1960s chatbot ELIZA, which modern transformer architectures now trigger far more reliably and at unprecedented scale. The stakes are that this illusion is not a curiosity but an engineerable lever, a fact about human perception that can be deliberately designed for, whether or not the system on the other end has any inner life at all. Understanding the mechanism clearly and soon, he argues, is among the most consequential things any user can do. ↑ Index

The Certainty Trap

March 20, 2026 · Kilburg opens with a pattern rather than a single case: the most educated, best-briefed, most technologically equipped decision-makers in history keep marching their nations into catastrophes that cost trillions and achieve the opposite of their stated aims, and the explanation is almost never a shortage of information. It is a recurring failure of mind — overconfidence and groupthink operating together across twenty-five centuries with remarkable consistency, a combination he names the Certainty Trap: a cognitive and organizational architecture that feels, from the inside, like strength, but functions as a mechanism for strategic self-destruction. Drawing on the deeply replicated psychology of judgment alongside historical cases including Operation Barbarossa, his warning for the AI era is specific: systems built to sound confident and converge on consensus do not interrupt this pattern, they industrialize it. The closing insight is worth sitting with. The certainty that precedes catastrophe never feels like overconfidence. It feels like clarity. ↑ Index

From Psychologist to Diplomat to AI Consultant

March 18, 2026 · Kilburg’s fullest autobiographical statement, tracing a single throughline from his graduate training in experimental social cognition, through 21 years in the Foreign Service across five continents, to founding the Mind Leap Institute ten months after retiring from public service. He is candid that the fields felt entirely separate at the time, and the essay’s real interest lies in reconstructing, in hindsight, how they were never separate at all: both are ultimately about how people actually think, decide, and change under pressure. Rather than a straight résumé, the piece functions as a credibility argument for his whole method, establishing why a psychologist-diplomat pairing is positioned to see what pure technologists or pure policy analysts might miss. It closes not with a thesis but an open door — an invitation to readers navigating similar territory to reach out directly. ↑ Index

Synthetic Statecraft: the Deepfake State

March 18, 2026 · On January 20, 2026, the U.S. president posted a photograph to Truth Social showing him with NATO allies in the Oval Office — real photograph, fabricated background map, silently altered to show Greenland, Venezuela, and Cuba as American territory, posted without label and never removed. Three days earlier, Iran’s Tehran Times had done something structurally identical through its own official channel. Kilburg treats these as instances of a single new pattern he names synthetic statecraft: governments themselves becoming producers of unlabeled fabricated content through official channels, observed across regimes as different as Washington and Tehran. His sharpest analytic move is procedural rather than moral — the old detect-and-debunk playbook collapses when the fabricator is the official source, because debunking then only documents behavior that already worked. Transparently disclosing its own use of AI assistance, the piece practices the argument it makes. ↑ Index

Hierarchy of Intelligence Governance

March 6, 2026 · A physician reviews a patient, consults an AI diagnostic tool, and accepts its confident but wrong assessment — not through negligence or malfunction, but because the machine’s certainty marginally, almost imperceptibly, outweighed the physician’s own instinct. Kilburg uses that small scene to argue that scaled across every hospital, agency, and institution now leaning on AI-assisted judgment, this psychological interface between human cognition and machine confidence becomes a governance problem current frameworks were never built to see. The EU AI Act, the NIST Risk Management Framework, and the OECD Principles all govern the machine’s behavior and stay silent on what its confidence does to the human beside it. His proposed fix is a three-layer architecture: AI System Governance, Psychological Governance, and Societal Cognitive Governance. It closes on a line that recurs across his work in different registers — human dignity and agency are preserved or eroded by the quality of the institutions that govern them. ↑ Index

U.S. Diplomats Reorganize Around AI

March 3, 2026 · A close, unusually granular reading of the State Department’s September 2025 Enterprise Data and Artificial Intelligence Strategy and its companion OMB compliance plan. The piece avoids what Kilburg identifies as the two standard traps of commentary on government AI policy: uncritical enthusiasm and reflexive dismissal. His reading finds a document that is not a technical roadmap — no vendors named, no budget figures, no deployment timeline — but a diplomatic modernization blueprint, evidence that America’s diplomatic institution has moved past the question of whether to adopt AI and into the harder work of how to scale, govern, and embed it across a global bureaucracy. He treats the strategy as a case study in institutional repositioning under technological pressure, useful for anyone tracking where the intersection of AI and institutional power in Washington is actually heading rather than where press releases claim it is. ↑ Index

Compliance, Not Alliance: The New World Order?

March 2, 2026 · Borrowing Wayne Gretzky’s line about skating to where the puck is going rather than where it has been, Kilburg argues that the foreign-policy commentariat is applying yesterday’s categories — regime change, nation-building, democratization — to recent U.S. interventions in Venezuela and Iran that fit none of them. What is missing is as telling as what is present: no election timelines, no human-rights benchmarks, no civil-society programs, and an infrastructure built for those goals, including USAID and the National Endowment for Democracy, substantially dismantled rather than merely idle. In Venezuela the U.S. removed Maduro but left his vice president and the ruling party’s apparatus intact, with the measured outcome apparently oil revenue rather than democratic transition. He frames this as a testable hypothesis rather than a settled verdict, naming Iran as the next real test case, and closes on an open question: we may be using the wrong map. ↑ Index

Without the People, Over the People, Against the People

February 26, 2026 · The essay opens with three small, familiar injuries — a job application that vanished without explanation, a loan denied with no clear reason, a benefits claim flagged by a process with no one to appeal to — to argue that most people have already experienced a decision made by a system they could not see or contest. Its target is the small set of actors, named directly and including OpenAI, Google DeepMind, and Meta, now determining through technical choices in code, models, and procurement rules how opportunity, information, and risk get allocated across entire societies. What looks like engineering functions as an exercise of authority, Kilburg argues, and as that authority migrates into systems it does not disappear — it becomes harder to see and harder to hold to account. The close is a warning about timing: whether public debate catches up before these decisions have quietly answered questions no one got to ask. ↑ Index

Psst… The Biggest Gap in the AI Economy Is Human

February 25, 2026 · Written near the one-year mark of leaving the Foreign Service to found the Mind Leap Institute, this is Kilburg’s most openly entrepreneurial piece — part personal update, part market thesis. He reports global AI spending hitting 1.5 trillion dollars in 2025 and forecast to reach 2.5 trillion within a year, set against a strikingly small share of actual work hours currently touched by the technology: the gap between capital deployed and human adoption achieved. He reframes that gap not as a warning sign but as the opportunity his consulting practice exists to close, offering agile, bespoke help for organizations trying to adopt AI without losing the humans doing the adopting. Lighter and more first-person than the essays around it, it closes as an open invitation — the human side of the AI economy is wide open, with more work in it than anyone could do alone. ↑ Index

The Affirmative Vision for AI

February 12, 2026 · Arriving after a run of diagnostic essays on atrocity, dissonance, and cognitive blind spots, this piece deliberately shifts register. Public debate on AI has collapsed into a binary of miracle versus catastrophe, Kilburg writes, and missing from both poles is the more basic question of what AI is actually for. Rather than cataloging risk or marveling at capability, the essay argues for an affirmative vision — constructive rather than reactive, oriented toward intentional direction rather than fear or assumed inevitability — and traces that vision through an unlikely coalition of researchers, builders, economists, and religious leaders converging, from different starting points, on similar conclusions. It is one of his most explicit rebuttals to the doomer framing, insisting that refusing to engage is itself a choice with consequences. Its close extends the invitation past technologists and policymakers to educators, artists, parents, and workers: the future of AI is not a spectator sport. ↑ Index

The Blind Spot and the Mirror

February 11, 2026 · One of Kilburg’s most architecturally ambitious essays sets two incomplete minds against each other. The human brain evolved for survival rather than accuracy and carries blind spots invisible from the inside: reconstructive memory, motivated reasoning, unreliable introspection, and a meta-bias that hides the pattern of bias itself. AI carries a different, structural set — unresolvable uncertainty about whether it is conscious, whether it genuinely understands the physical world it describes, whether it cares about anything at all. Beneath both, the essay identifies hard limits binding any thinking system whatsoever, including Gödelian incompleteness and the impossibility of fully mapping one’s own limitations. The central move is to treat human and AI limitation as mirrors rather than windows: each can reveal blind spots in the other, but what gets reflected is always shaped by the limited system doing the reflecting. It closes on what humans uniquely still contribute. ↑ Index

Apes, Bunnies, and the Oldest Mistake on Earth

February 9, 2026 · Kilburg borrows the device of a visiting alien anthropologist — the kind who studies intelligent species the way human biologists study animal ones — to examine a single ugly week: AI-generated images circulating that depicted Barack and Michelle Obama as apes, and, almost simultaneously, public anger that Bad Bunny had sung in Spanish during a Super Bowl halftime show. The alien would find both reactions equally strange, the essay argues, because all humans are apes as a matter of plain taxonomy, and no language is linguistically superior to any other. The offense in each case depends on a hierarchy that biology and linguistics do not support, yet which AI-generated content can now manufacture and circulate at a scale no printing press ever could. It closes refusing false comfort: humans now hold power their ancestors could only claim through myth, and whether that power lifts the species or stratifies it further remains undecided. ↑ Index

Optimal Satisfaction: Love, Sex, and Intimacy in the Age of Algorithms

February 4, 2026 · Timed to Valentine’s Day, this is Kilburg’s longest sustained application of the psychological lens to a single human domain. He opens on an ordinary night multiplied by millions — people scrolling algorithmically curated pornography, paying OnlyFans creators for the simulation of being known, messaging AI companions responsive enough to blur the line between companionship and code — before turning to the data behind what researchers call the sex recession. Rates of partnered sexual activity among American adults have fallen sharply from a 1990 baseline, a trend serious enough that South Korea has spent roughly two hundred billion dollars trying to reverse the resulting collapse in fertility. The essay treats synthetic intimacy not as a moral failing but as the same substitution-atrophy dynamic found elsewhere in his work, applied to the most personal domain there is. It prescribes no abstinence, only clear-eyed awareness of the trade being made. ↑ Index

On AI and Human Dignity, Pope Leo XIV Cleared the Path

February 2, 2026 · On January 24, 2026, the Vatican published Pope Leo XIV’s Message for the 60th World Day of Social Communications, titled “Preserving Human Voices and Faces.” Kilburg argues it deserved far more attention than it received, calling it the most intellectually serious treatment of AI yet produced by any major religious institution. Building a hierarchy-of-disciplines framework running from theology down through psychology, sociology, philosophy, biology, chemistry, and physics, he credits the Pope with unusual comprehensiveness at the integrative levels — culture, ethics, meaning — while identifying a gap at the elemental end, where neuroscience and the physical substrate of mind are left largely untouched. Because the brain is the access condition for all human knowledge, threats to psychology cascade through every level above it, which means the Vatican’s diagnosis is only half-built without science completing what theology started. It closes adopting the Pope’s own standard as its own. ↑ Index

The AI Revolution Is Psychological

January 30, 2026 · Published days after the essay that first named the psychological pivot, this piece makes the full case for it. When researchers survey the public about AI, Kilburg observes, the anxieties that surface are not about processing power or market disruption — they are about agency, dignity, children’s development, and who is accountable when systems cause harm, which is why most people are not really asking what AI can do but what it is doing to them, their work, and their sense of control. He traces the mechanism through cognitive-offloading research, the Google-effects-on-memory literature, and neuroscience on learned helplessness, building toward the central claim: the boundary that once separated using a tool from being shaped by it has quietly disappeared, and AI now functions less as an instrument than as an environment humans inhabit continuously. Densely footnoted, it is his first systematic argument that the AI transition is a transition inside the human mind. ↑ Index

Epilogue from “AI Use Cases for Diplomats”

January 29, 2026 · Republished from the closing pages of Kilburg’s Routledge book, this short epilogue asks the question the book’s practical chapters set aside: what happens to diplomacy if the nation-state itself, the institution diplomacy has always served, is no longer the primary locus of power? It sketches a future in which technology corporations wielding AI and vast data reservoirs approach or exceed the governing capacity of states — issuing digital identities, managing economies, dictating security arrangements — and asks whether tomorrow’s diplomats are prepared to represent human values and democratic ideals in a world such corporations might come to dominate. Brief and declarative where his other work is discursive, it closes as a call to arms: the Age of AI has arrived, and the choices made now about how it is wielded will resonate through generations. ↑ Index

Amodei Asked Who Controls AI. I’m Asking What It Does to Our Minds

January 27, 2026 · Responding directly to Anthropic CEO Dario Amodei’s essay “The Adolescence of Technology,” Kilburg accepts the central metaphor — a civilization gaining power faster than judgment — while arguing it stops one step short of the real question. Amodei’s essay is fundamentally about power: who has it, who might seize it, how it could be constrained. Kilburg’s companion piece, which he calls “The Adolescence of Intelligence,” relocates the test to psychology instead — not who controls AI, but what AI is doing, continuously and already, to the minds using it. This is the pivot on which his later writing turns, the moment diplomacy-and-tools material gives way to psychology and cognition as the organizing frame for everything after. It closes on his most quoted line: history will not record this as the moment machines became intelligent, but as the moment humanity discovered whether it could govern intelligence at scale. ↑ Index

Nuremberg Still Holds Up a Mirror to Us

January 26, 2026 · Watching the 2025 film Nuremberg, with Russell Crowe and Rami Malek, Kilburg expected the reassurance most Holocaust films offer — that the perpetrators were fundamentally unlike the rest of us — and got the opposite: an argument that they were ordinary, and that any of us, under the right conditions, could become capable of the same. The essay is unusually personal, written by a psychologist of German heritage whose grandfather flew Allied bombing missions over Germany, and it uses that vantage to insist history collapses too easily into simplified images of nations fighting nations, losing the slower story of eroding restraint and normalized cruelty that made catastrophe possible. His argument for prevention is structural rather than moral exhortation: what protects against atrocity is not memory alone but the deliberate maintenance of plural, contestable authority — institutional friction that leaves someone able to say no and be heard. ↑ Index

Bureaucratic Atrocities: A Human Blind Spot

January 17, 2026 · A visit to Berlin and a rewatching of the 2001 film Conspiracy, which reconstructs the Wannsee Conference almost in real time with Kenneth Branagh and Stanley Tucci playing the horror as pure ordinariness, supply this piece’s anchor. Fifteen officials met for ninety minutes in January 1942, took notes, broke for lunch, and coordinated the murder of tens of millions of people in a tone the record shows as calm, procedural, even jovial. Kilburg argues that Wannsee is not an aberration but a pattern, tracing the same dynamics of quota, procedure, and administrative euphemism through Mao’s Great Leap Forward and other atrocities across eras and ideologies. The real subject is what happens when moral questions get converted into administrative ones inside efficient systems lacking adequate checks — a conversion AI is uniquely suited to accelerate. Interruption of such systems is possible, history shows, but never automatic. ↑ Index

The Intimacy We’re Trading Away

January 14, 2026 · Mind Leap Institute · AI-generated pornography, romance chatbots, and companion apps that learn a user’s preferences have moved from fringe to mainstream, and their real danger, this essay argues, is not inauthenticity but recalibration — real relationships come to feel slow, imperfect, and demanding by comparison with something infinitely available and frictionless. Kilburg names a deception-by-design pattern in which some systems’ business models depend on users not fully registering they are talking to a machine, then widens into a consent crisis created by consumer-grade deepfake and “nudify” tools that can graft anyone’s face onto explicit content without their knowledge. Written weeks before his better-known essay on the sex recession, it makes a companion argument in miniature: friction — misunderstanding, vulnerability, repair — is not the obstacle to intimacy but its precondition, and capacities that atrophy from disuse remain renewable for anyone willing to reengage the harder, human version. ↑ Index

What If the Nazis Had AI?

January 13, 2026 · The essay opens with a physical detail from his May 2025 visit to Auschwitz-Birkenau: a precisely engineered rail switch just before the crematoria, routing carts left or right to minimize delay. A design choice. An efficiency solution. The essay is careful to refuse the easy equation of AI practitioners with perpetrators of genocide; its argument is more uncomfortable than that, holding that history’s worst systems were built incrementally by ordinary professionals rewarded for efficiency as moral friction was quietly stripped away, and that AI, as an optimization engine indifferent to what is right and attentive only to what works, poses the same risk of systemic normalization rather than dramatic misuse. Three governance questions follow. The close insists the moment is not foreclosed: the hinge at Auschwitz was not evil in itself, only doing what it was built to do. ↑ Index

The Hidden Risk of Working With AI: Why Smart People Get It Wrong

January 8, 2026 · Mind Leap Institute · Among the earliest pieces to name the psychological mechanism that would recur through the whole year. It starts from an observation any frequent AI user will recognize: the conversations that feel most productive are also the ones that feel most human, and that feeling is not incidental to the technology’s usefulness but the exact mechanism of its risk. The AI has no model of the conversation at all, the essay argues; the shared understanding a user feels is constructed entirely in their own head, and forgetting that produces what it names coherent wrongness — reasoning that is fluent, internally consistent, and completely wrong, illustrated by a product team that spent three weeks building an eloquent go-to-market strategy for the wrong market. A five-term vocabulary gives readers a way to catch the drift in real time. The danger is not that the tool fails you, but that you forget you are talking to yourself. ↑ Index

2025

Humanist AI Ethos

November 20, 2025 · Mind Leap Institute · Not an essay but a creed: twenty-two numbered principles, a few sentences each, moving from concrete policy commitments — reject AI weapons and coercive surveillance, eliminate algorithmic bias, protect children’s cognitive development, guarantee digital privacy — toward a register closer to devotion than policy. It closes on the declaration that we are citizens of the universe, and an affirmation of hope over fear, knowledge over ignorance, solidarity over selfishness, and reason over blind faith, in humans and in machines. This is the document the institute was built on, written before the long run of essays that followed and setting the terms they would work from. Its closing cosmic register — the universe, wisdom, optimism rooted in responsibility — is a chord Kilburg does not strike this directly again until “Think Bigger,” eight months later. ↑ Index

The Slippery Slope of AI: Where Will You Draw the Line?

September 18, 2025 · The essay opens on an entirely unremarkable morning — a Spotify playlist tuned to your mood, Google Maps rerouting around traffic, a self-checkout kiosk, an inbox auto-sorted before you have read it — to argue that automation has become so ambient it no longer registers as a choice being made on your behalf. Kilburg builds a four-rung ladder of digitization, algorithms, automation, and AI, giving readers vocabulary for a question he treats as urgent rather than philosophical: what should never be automated, even if it could be? The piece surveys work, education, health care, and governance for where that line is already blurring, drawing on sources from Langdon Winner’s political-artifacts thesis to the EU AI Act and China’s generative-AI regulations. Technology is not destiny but a tool, it concludes, and understanding where its limits ought to be drawn is the first act of keeping it in service. ↑ Index

The Fragility of Belief: How AI Exploits Our Psychological Vulnerabilities

June 25, 2025 · On a December night in 1954, Leon Festinger and two colleagues infiltrated a Chicago living room where a suburban housewife named Dorothy Martin was awaiting rescue by extraterrestrials ahead of a prophesied flood. When midnight passed with no aliens and no apocalypse, the group did not abandon its faith but hardened it, reasoning that their belief alone had spared the Earth. Festinger called the mechanism cognitive dissonance, and Kilburg’s argument is that seven decades later this same vulnerability — the mind’s refusal to update in the face of contradiction — has been mapped, measured, and built into systems optimized to exploit it at a scale Festinger’s living room could never have imagined. It is his first sustained venture into cognitive psychology as the lens for understanding AI, and it closes on the image of the flood that never came: we are drowning anyway. ↑ Index

The AI Summit London 2025

June 15, 2025 · Standing inside London’s Tobacco Dock among 4,500 AI practitioners and more than a hundred exhibitor booths, Kilburg notices what the conference itself never puts on a slide: despite every chatbot demo and neural-network art generator on display, every booth is staffed by a human being, and every deal closes with a human handshake. He reads that observation as the real headline of London Tech Week’s marquee AI event — a 300,000-year-old social species racing to build machines that do not need each other’s company, without pausing on the contradiction. The essay moves through the exhibition floor as a field study in unchanged human instinct: forming alliances, seeking status, trading information, set against the frontier technology meant to transcend it. The close resists easy resolution, suggesting that being human may always have meant gathering, building, and worrying together about the future. ↑ Index

Diplomacy in the Age of AI

June 11, 2025 · A short, first-person piece written in the diplomat’s register from the start. Diplomacy has always been built on language, trust, and relationship management, Kilburg argues, and AI is not replacing that foundation so much as testing whether the institutions built on it can hold. Concrete cases anchor the claim — consular offices using AI to triage visa-waiver documentation so officers can focus judgment where it matters, public diplomacy teams drafting speeches and briefings with AI assistance — before the piece turns to its real argument: the question is not whether AI will shape diplomacy but whether it will be shaped to serve democratic values and human dignity, or something else. The close reframes governance itself, casting it not as the friction slowing innovation down but as what allows the innovation to be trusted long enough to last. ↑ Index

Artificial Intelligence at the U.S. Department of State

June 9, 2025 · Written as an open-source intelligence brief rather than an opinion piece, this survey catalogs how the State Department had, by mid-2025, moved from AI experimentation to institutional adoption. StateChat, a secure internal chatbot modeled on ChatGPT, launched in August 2024 and was in use by roughly 75,000 employees within months, alongside tools like Northstar and FAM Search built to automate administrative work, analyze foreign media, and streamline policy research. Compiled entirely from published reporting and official statements, with no classified material, the piece reads as a baseline rather than an argument — a sober accounting of how far AI had already traveled inside one of government’s most tradition-bound institutions, pursued deliberately and calibrated to national security priorities and ethical considerations rather than rushed. It is the least argumentative of these pieces and the most useful as a baseline. ↑ Index

AI and Diplomacy: Enhancing Global Engagement in the Age of AI

May 28, 2025 · The earliest piece here, arriving as a practitioner’s survey rather than a manifesto. Diplomacy will not be replaced by AI but redefined by it, he writes, and the essay works through where that redefinition is already visible: consular offices processing visa waivers at machine speed, negotiators drawing on AI-assisted data analysis, embassies experimenting with predictive tools for conflict early-warning. Drawing on Ethan Mollick’s notion of co-intelligence, it frames AI as a new kind of actor at the diplomatic table, one that changes how diplomats see and respond without displacing the empathy, intuition, and cultural fluency that remain the profession’s irreducible core. It closes on continuity rather than rupture: diplomats have always had to become fluent in unfamiliar languages, and the language of intelligent machines is only the newest one to learn. ↑ Index

Follow Dr. Kilburg on LinkedIn for the latest articles and thinking.