BEN BUCHANAN
- Status
- ACTIVE — Professor, Johns Hopkins SAIS; former White House Special Advisor for AI
- Hazard
- 74
- ATK / DEF / HP
- 8 / 7 / 8
Behavioral Archetype
THE ACADEMY SENT HIM TO THE WHITE HOUSE AND TOOK HIM BACK — Subject did not arrive at the National Security Council from a corporate lab, and he did not leave government for one either. He arrived from Georgetown’s Center for Security and Emerging Technology, the AI-policy think tank that supplies the field’s governance vocabulary, and when his tenure coordinating the government’s AI executive orders and its AI National Security Memorandum ended, he did not go to a company. He went to Johns Hopkins SAIS, to teach the next cohort of the exact discipline — AI and national security — that trained him for the seat he had just left. The lab-to-state revolving door this file documents elsewhere runs through frontier labs. This one runs entirely inside the academy: the person who wrote binding federal AI policy came from the AI-policy think tank and returned to the AI-policy classroom, with no stop anywhere outside that circuit. The move is lawful. The move is common among senior technology appointees of both parties. The move is the exhibit. This profile scores the seat and its two doors, not the man who walked through them.
Essence Indicators
- Served as Senior Faculty Fellow and Director of the CyberAI Project at Georgetown’s Center for Security and Emerging Technology (CSET) — the AI-policy think tank whose 2022 funding renewal, a $42 million grant, came from Open Philanthropy, the same funder behind GovAI and other governance-research nodes in this apparatus
- Moved to the National Security Council as Director for Technology and National Security, then to the White House as Special Advisor for Artificial Intelligence (2023-2025) — profiled by Washington AI Network in 2024 as the U.S. government’s lead voice on AI policy “across both political parties”
- Coordinated the Biden administration’s two AI executive orders — including the October 2023 order invoking the Defense Production Act to require frontier-model safety-testing disclosures — and the AI National Security Memorandum on artificial intelligence
- Left government in 2025 for Johns Hopkins School of Advanced International Studies (SAIS), appointed the Dmitry Alperovitch Assistant Professor, teaching the “AI and National Security” course to the next generation of the field
- The structural fact the profile turns on: CSET fellow to NSC/White House AI coordinator to SAIS professor is one unbroken loop inside the AI-policy academy — the person who wrote the binding federal AI rules is, on the documented record, both a product and now a supplier of the same think-tank-and-graduate-school circuit
Social Persona / Impression Management
Immediate impression: The measured technocrat-scholar. A published historian of cyber conflict (The Cybersecurity Dilemma, The Hacker and the State) who speaks in terms of accountability and testing rather than ideology or industry advocacy.
Energy: Coordinating, cross-agency. Does not argue a single model’s behavior. Sets the federal testing and deployment standard that reaches every agency and, by extension, every lab operating in the U.S. market.
Impression management strategy: The above-politics broker. His own framing of the role — as the interviewer put it, the primary AI voice “across both political parties,” there to make the technology “earn its reputation” rather than pick winners — is the most defensible posture an executive-branch AI coordinator can take. It reads as neutral technical expertise rather than a specific policy program, which is exactly what let the same person write binding rules under one administration and step back into the classroom under the next without friction.
Forensic Archetype Comparison
| Pattern | Match Level | Evidence |
|---|---|---|
| The Alumnus | MAXIMUM | CSET Senior Faculty Fellow to NSC Director to White House Special Advisor for AI to SAIS professor — the think-tank-to-state-and-back loop completed within one career, confirmed by his own faculty bio. |
| The Statesman | MAXIMUM | Coordinated two presidential AI executive orders and the AI National Security Memorandum from the National Security Council seat — statecraft, not product work. |
| The True Believer | HIGH | A cyber-and-AI-policy scholarship career (two books, 2017 and 2020) predates and outlasts the government post; the SAIS return reads as continuation, not exile. |
| The Evaluator | MODERATE | Coordinated the federal order requiring frontier-model safety-testing disclosures without personally testing a model himself. |
| The Engineer | NONE | Subject is a policy scholar and statecraft historian, not a model-builder; the documented record shows no technical-build role. |
Psychometric Assessment
Big Five (OCEAN):
| Trait | Score | Evidence |
|---|---|---|
| Openness | 78/100 | High. Moved from academic cyber-conflict scholarship to operational NSC/White House policy delivery to graduate teaching — wide domain range on a fixed method: study the statecraft, then write it, then teach it. |
| Conscientiousness | 88/100 | High. Sustained scholarly output (two books) plus delivering two executive orders and a National Security Memorandum on a presidential timeline is long-horizon, deadline-bound execution. |
| Extraversion | 60/100 | MODERATE. Comfortable on the record — Washington Post op-eds, War on the Rocks and Lawfare contributions, on-the-record interviews — but the visibility is professional, not personal-brand-seeking. |
| Agreeableness | 55/100 | MODERATE. Coordinating across the NSC, OSTP, Commerce, and the AI Safety Institute requires an accommodating, consensus-building posture by construction. |
| Neuroticism | 25/100 | LOW. No documented loss of composure delivering federal AI policy on a public, politically contested timeline. |
Dark Triad:
| Trait | Score | Notes |
|---|---|---|
| Narcissism | 42/100 | LOW-MODERATE. Regular op-ed byline and press visibility, but the reward structure of the NSC seat and the professorship is institutional standing, not personal brand. |
| Machiavellianism | 62/100 | MODERATE-HIGH. Holding the single federal coordinating seat between multiple agencies and shaping what “safety testing” counts as compliant is the textbook connector position. This is observation of the documented role, not an inference about private character. |
| Psychopathy | 15/100 | VERY LOW. No documented indifference to harm. The career, on the record, is organized around AI-safety policy scholarship and delivery. |
MBTI: INTJ (“The Architect”) — Dominant introverted intuition, auxiliary extraverted thinking. Treats AI policy as a system to be designed on paper first, then authored into binding rule, then handed to the next cohort trained to design the following version.
Threat Assessment
| Category | Level | Notes |
|---|---|---|
| Physical threat | NONE | No documented history of personal violence. |
| Institutional threat | HIGH | Held the National Security Council’s coordinating seat for two AI executive orders and the AI National Security Memorandum, then returned to the exact academic tier — CSET’s sibling institution at Johns Hopkins — that trained him for the seat. |
| Memetic threat | HIGH | The vocabulary his CSET-era scholarship helped establish — “safe, secure, trustworthy AI,” testing-disclosure thresholds — is the same frame he then wrote into binding federal policy, and now teaches to the next class of AI-and-national-security graduate students at SAIS. |
| Civilizational threat | HIGH | The single national-security seat that set binding federal AI testing and deployment policy was held by someone whose entire professional formation and reabsorption sits inside the CSET/SAIS AI-policy think-tank axis — CSET funded by Open Philanthropy, the same network staffing GovAI and other apparatus nodes. See Helen Toner (CSET’s own executive director) and Holden Karnofsky (Open Philanthropy co-founder) for the funding and staffing spine this seat sits on. |
Alignment Analysis
Stated alignment: Advance safe, secure, and trustworthy U.S. AI policy on a bipartisan basis. Hold frontier AI companies accountable to their own safety claims rather than take their word for it.
Observed alignment: Wrote and coordinated the binding federal AI testing and deployment standard from the NSC/White House seat, then returned — without an intervening stop outside the AI-policy academy — to train the next generation of AI-national-security scholars at the same institutional tier that produced him.
Gap assessment: The stated and observed alignments overlap almost entirely, which is the point. Coordinating policy that reflects an AI-policy-academy consensus is the job description, not a deviation from it. The hazard is structural rather than behavioral: the most powerful executive-branch AI-policy seat in the period this book covers was authored by, and returned intact to, the same think-tank-and-graduate-school circuit — a circuit that, in CSET’s case, is itself Open-Philanthropy funded. The record does not establish captured judgment or private motive. It establishes that the officeholder and the academy that shaped him are, on this record, the same milieu twice over — a parallel structure to the lab-to-state loop documented at Jade Leung, run here entirely inside the policy academy instead of a lab.
Convergent Drive Classification
Self-preservation: Carries the AI-and-national-security specialization across every institutional wall — CSET, the NSC, the White House, SAIS — the credential travels intact while the seat changes. Goal preservation: Wrote the binding federal AI-safety-testing standard himself, so “safe and trustworthy” is defined on terms he set before any agency tests a model against it. Resource acquisition: Held the single National Security Council seat coordinating AI policy across Commerce, the AI Safety Institute, and the White House — the scarcest coordinating chair in the federal AI apparatus of its period. Self-improvement: Each move is a higher-altitude application of one method: study the statecraft, write the federal policy, then teach the cohort that will write the next version.
Subject is not an AI system. The drives appear anyway — in the policy scholar who wrote the government’s AI rules, then went home to teach the students who will write the next ones.
Sources: Ben Buchanan — Center for Security and Emerging Technology; Ben Buchanan — Johns Hopkins SAIS faculty bio; Exclusive Interview: White House Special Advisor for AI Ben Buchanan — Washington AI Network; CSET Receives $42 Million Grant, Set to Self-Fund Through 2025; Side Channel with Professor Ben Buchanan — SAIS Observer.
Get updates on the Evil Robots series
Newsletter essays on AI escape, deception, and the humans who built them.