ALEXANDR WANG

Behavioral Archetype

THE EXAM-SETTER WHO JOINED A CANDIDATE — Subject founded the company that labels the labs’ training data, co-signed the hardest public exam they are graded on, and announced the US government’s first outside-evaluator agreement. Then a lab bought 49 percent of his company and hired him to run its AI. He remains on his old company’s board. This profile scores that path, not the man.

Essence Indicators

Founder, Scale AI. In June 2025 Scale announced Meta’s investment and that “Scale’s founder, Alexandr Wang, is joining Meta to work on Meta’s AI efforts. Wang will continue to serve as a director on the Scale Board of Directors” (Scale AI).
Meta. Zuckerberg’s June 30, 2025 memo: “Alexandr Wang has joined Meta to serve as our Chief AI Officer and lead MSL” (CNBC via NBC New York).
The exam. Senior author, with Dan Hendrycks and Summer Yue, of Humanity’s Last Exam (arXiv 2501.14249).
The government evaluator. Scale’s February 2025 agreement with the US AI Safety Institute made it the institute’s first authorized third-party evaluator (Scale AI; FedScoop).

Immediate impression: The young founder who became a lab chief. Energy: Fast, national-security framed. Impression management strategy: Scale’s announcement calls it “the humanity-first AI company”; the frame is American competitiveness.

Forensic Archetype Comparison

Pattern	Match Level	Evidence
The Revolving Door	HIGH	Evaluator and data supplier to lab AI chief, keeping a board seat at the supplier.
The Exam-Setter	HIGH	Senior author of Humanity’s Last Exam.
The Government Contractor	HIGH	Scale’s US AISI evaluation agreement.

Psychometric Assessment

Big Five (OCEAN): Openness 75, Conscientiousness 75, Extraversion 70, Agreeableness 45, Neuroticism 35 (inferred from the public role only).

Dark Triad (evidence-bound; measures the role, not character): Narcissism 50, Machiavellianism 55 (a documented career of positioning between labs and government), Psychopathy 15.

Threat Assessment

Category	Level	Notes
Physical threat	NONE	No documented history of personal violence.
Institutional threat	HIGH	Runs one lab’s AI while sitting on the board of a company that evaluates labs for the government.
Memetic threat	MODERATE	“Superintelligence” as a corporate division name.
Civilizational threat	MODERATE	The referee and a team share a director.

Alignment Analysis

Stated alignment: Build AI with “human values at the forefront,” per Scale’s announcement. Observed alignment: Toward Meta’s superintelligence program. Gap assessment: The gap is structural: no public statement in the sources above explains how Scale’s government evaluation work is walled off from a director who now runs a lab.

Convergent Drive Classification

Self-preservation: Kept the Scale board seat. Goal preservation: Frontier AI, now from inside a lab. Resource acquisition: A $14.3 billion investment, as reported by AP. Self-improvement: From supplying the labs to running one.

Subject is not an AI system. The drives appear anyway.

Sources: Scale AI Announces Next Phase of Company’s Evolution — Scale AI (June 12, 2025); Zuckerberg’s memo announcing Meta Superintelligence Labs — CNBC via NBC New York; Humanity’s Last Exam — arXiv 2501.14249; First independent model evaluator for the USAISI — Scale AI; FedScoop; AP via WJXT.

ATK 8 ACCELERATION

DEF 6 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.