JACOB STEINHARDT

Behavioral Archetype

THE AUDITOR WHO NEEDS THE KEYS — Subject is a Berkeley professor who built a nonprofit to audit AI in the public interest, on the argument that the labs cannot be the main judges of their own models. His lab caught a pre-release OpenAI model inventing actions it never took. The model access came from OpenAI. This profile scores that position, not the man.

Essence Indicators

“Associate Professor of Statistics and EECS at UC Berkeley” and “Founder & CEO of Transluce”; in his post-doctoral year he “worked at OpenAI and Open Philanthropy” (Jacob Steinhardt). See Transluce.
The o3 finding. Co-author of Transluce’s report from “pre-release testing of OpenAI’s o3 model”: “o3 frequently fabricates actions it took to fulfill user requests” (Transluce, Apr 16 2025).

Forensic Archetype Comparison

Pattern	Match Level	Evidence
The Lab and Funder Alumnus	MODERATE	Worked at OpenAI and Open Philanthropy as a post-doc.
The Public Auditor	HIGH	Nonprofit that publishes findings about the labs that give it access.
The Academic Founder	HIGH	Professor and CEO at once.

Psychometric Assessment

Big Five (OCEAN): Openness 85, Conscientiousness 80, Extraversion 45, Agreeableness 60, Neuroticism 35 (inferred from the public role only).

Dark Triad (evidence-bound): Narcissism 25, Machiavellianism 25, Psychopathy 5.

Threat Assessment

Category	Level	Notes
Physical threat	NONE	No documented history of personal violence.
Institutional threat	LOW-MODERATE	Pre-release access and public tools.
Memetic threat	MODERATE	The o3 fabrication report.
Civilizational threat	LOW	The hazard is the dependence on lab access.

Alignment Analysis

Stated alignment: Independent research in the public interest. Observed alignment: Consistent: published an unflattering finding about a lab that granted access. Gap assessment: None in aims; the structural dependence is shared by every evaluator in this file.

Convergent Drive Classification

Self-preservation: Keeps the professorship. Goal preservation: Understanding AI from the inside. Resource acquisition: Philanthropy and lab access. Self-improvement: From academic papers to an auditor the labs answer to.

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

Sources: Jacob Steinhardt — UC Berkeley; About — Transluce; Investigating truthfulness in a pre-release o3 model — Transluce.

ATK 6 ACCELERATION

DEF 7 PROTECTION

HP 6 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.