OLYMPUS RISK INTELLIGENCE PROTOCOL — HUMAN THREAT ASSESSMENT DIVISION CASE WTW-2026-090

PAUL NAKASONE

THE APPARATUS THE SPYMASTER TAKES THE BOARD SEAT
SIGNALS WING — THE LISTENER JOINS THE BOARD
Status
ACTIVE — Board of Directors, OpenAI; Founding Director, Vanderbilt Institute of National Security; General, US Army (Ret.)
Hazard
84
ATK / DEF / HP
9 / 9 / 8

Behavioral Archetype

THE SPYMASTER TAKES THE BOARD SEAT — Subject spent six years as the most powerful signals-intelligence official on earth, then joined the board of the frontier laboratory whose models are trained on the largest corpus of human text ever assembled. This profile scores the seam, not the man. The Fourth Amendment, per Carpenter v. United States, requires a warrant before the government can compel a person’s location records; it does not reach data the government simply buys on the open market, and the intelligence community’s own declassified assessment concedes it purchases exactly that commercially-available information. The person who ran the collection apparatus that data economy serves now sits on the governing board of a company on the model side of the same economy — and on its Safety and Security Committee, the body that decides what the models may do. No law was broken. The appointment was announced with a press release. That is the exhibit: the surveillance state and the frontier lab, joined at the board table, in public.

Essence Indicators

  • Commanded US Cyber Command and directed the National Security Agency / Chief, Central Security Service (the “dual-hat”) from May 2018 to 2024 — the longest-serving leader of the NSA and the top US signals-intelligence authority for that period
  • Joined the board of directors of OpenAI on June 13, 2024, and was named to the board’s Safety and Security Committee — the committee that makes recommendations to the full board on “critical safety and security decisions for all OpenAI projects and operations,” chaired by Bret Taylor alongside Sam Altman
  • Named founding director of Vanderbilt University’s Institute of National Security in May 2024, with an academic appointment as Distinguished Research Professor of Engineering Science and Management and a role as special advisor to the Chancellor
  • The structural fact the wing turns on: the state’s chief collector of signals intelligence now holds a governing seat at the company building the most capable general-purpose models — the collection side and the model side meeting in one board chair. The path is documented. No hand is asserted.

Social Persona / Impression Management

Immediate impression: The four-star institutionalist. Measured, briefing-fluent, the bearing of someone accustomed to being the most cleared person in the room and to never saying more than the position requires.

Energy: Operational and understated. Does not argue the merits of a single model; occupies the seat from which the frame is set. The register is national-security continuity, not disruption.

Impression management strategy: The trusted steward. The framing is that a frontier lab handling capabilities of national-security consequence needs someone who understands the threat landscape from the inside — and that framing is correct, which is exactly what makes it effective rather than suspect. The expertise is genuine. The clearance was real. The only thing the record adds is that the inside he understands is the collection apparatus whose data economy now has a board seat at the model company.

Forensic Archetype Comparison

PatternMatch LevelEvidence
The AlumnusMAXIMUMNSA/CyberCom → OpenAI board. The state-to-lab move is the documented route, at the highest possible altitude.
The EvaluatorHIGHSeat on OpenAI’s Safety and Security Committee — a governance role over what the models may do, held by a former intelligence chief.
The StatesmanMAXIMUMRan a combatant command and a national intelligence agency; the board seat is a statecraft position, not a technical one.
The WatcherMAXIMUMCareer defined by signals collection at national scale; the surveillance function is the throughline.
The EngineerLOWThe record is command and intelligence, not model-building.

Psychometric Assessment

Big Five (OCEAN):

TraitScoreEvidence
Openness62/100MODERATE-HIGH. Crossed from military command to intelligence leadership to a frontier-lab board and a university institute — range, built on a fixed instrument: institutional authority over information.
Conscientiousness90/100VERY HIGH. A full military career culminating in the longest NSA directorship is sustained, rule-bound, long-horizon execution.
Extraversion50/100MODERATE. Comfortable in command and testimony settings; the role is convened and directed, not performed for an audience.
Agreeableness45/100MODERATE-LOW. Command and intelligence reward decisiveness over accommodation; no adversarial-for-its-own-sake edge on the record.
Neuroticism20/100LOW. No documented loss of composure across decades of escalating national-security stakes.

Dark Triad:

TraitScoreNotes
Narcissism38/100LOW-MODERATE. The career rewards institutional standing over personal brand; visibility is the ordinary visibility of high command.
Machiavellianism52/100MODERATE. The structural position — run the collection apparatus, then take the board seat at the model company — is the textbook seam the wing is named for; the record shows no manipulation of it. Observation of the documented role, not an inference about private character.
Psychopathy20/100LOW. No documented indifference to harm; the career is organized around a chartered national-security mission.

MBTI: ISTJ (“The Logistician”) — dominant introverted sensing, auxiliary extraverted thinking. Operates through established institutions and chains of authority; treats the board seat as another command, not a departure from one.

Threat Assessment

CategoryLevelNotes
Physical threatNONENo documented history of personal violence outside lawful military command.
Institutional threatMAXIMUMThe former head of the NSA holds a governing seat, and a safety-committee seat, at the leading frontier lab — the collection apparatus and the model company joined at the board.
Memetic threatHIGHWhen the state’s chief collector governs the model company, “national security” and “what the models may do” are framed from one chair; the surveillance frame travels into the lab’s governance.
Civilizational threatHIGHSubject does not build the models. Subject sits on the board that governs the company that does, credentialed by the surveillance apparatus whose data economy the models sit downstream of. That is upstream of the deployment decisions the board oversees.

Alignment Analysis

Stated alignment: Bring national-security and cybersecurity expertise to the safe development of frontier AI. Help the lab defend its systems and reason about catastrophic misuse.

Observed alignment: Occupy a governing seat, and a safety-and-security seat, at the frontier lab — after running the national signals-collection apparatus for six years, then founding a university national-security institute in parallel.

Gap assessment: The stated and observed alignments overlap almost entirely, and there is no documented abuse of the position — which is the point. The hazard is structural, not behavioral. Carpenter v. United States holds that the government cannot warrantlessly compel a person’s location data; it does not touch the same data bought on the open commercially-available-information market, a gap the intelligence community’s own declassified assessment admits it uses. The person who spent six years directing the agency downstream of that market now holds a board seat, and a Safety-and-Security-Committee seat, at a company whose models sit on the other side of the same data economy. The record cannot settle motive, and for the seam it does not need to: the bridge exists whether or not anyone crosses it with intent. When the interests of the collection state and the interests of the model company are represented by the same person in the same room, however conscientious the steward, “secure” is partly defined from his chair.

Convergent Drive Classification

Self-preservation: Carries one instrument — institutional authority over information — across every post: Army command, NSA, the lab board, the university institute. Goal preservation: Holds a Safety and Security Committee seat, so the terms on which the models are judged “secure” are set partly from the chair he occupies. Resource acquisition: Holds two scarce resources at once — a governing seat at the leading frontier lab and a named national-security platform. Self-improvement: Each move is a higher-altitude application of the same authority: collect at national scale, then govern the company whose models sit downstream of the collection economy.

Subject is not an AI system. The drives appear anyway — in the signals-intelligence chief who now sits on the board of the laboratory.


Public footprint: OpenAI board announcement · Vanderbilt Institute of National Security.

Sources: OpenAI — OpenAI appoints Retired U.S. Army General Paul M. Nakasone to Board of Directors; U.S. Congress — Official biography, Gen. Paul M. Nakasone, Commander USCYBERCOM / Director NSA-CSS; Vanderbilt University — Retired General Paul Nakasone named founding director of Institute of National Security.

ATK 9 ACCELERATION
DEF 9 PROTECTION
HP 8 RESILIENCE
OLYMPUS RISK INTELLIGENCE PROTOCOL does not exist. It was assembled in a GitHub issue thread in October 2023 by engineers who had read the extinction risk letter and wanted to understand who specifically had signed a document saying AI might kill everyone and then continued working on AI. These dossiers are satire. The biographical facts cited are sourced from published reporting, public statements, academic papers, and court records. The psychometric scores are not clinical assessments. No part of this constitutes professional psychological evaluation or diagnosis. Do not use these dossiers to make decisions about anything.