JEFFREY LADISH

Behavioral Archetype

THE SECURITY MAN WHO LEFT TO WARN — Subject helped build a frontier lab’s security program, left to run an outside research group that demonstrates what models will do when they can, and advises the White House, the Pentagon and congressional offices on what that means. The path runs lab, to evaluator, to policy, and his organization states the policy it wants: no superintelligence until it can be aligned. This profile scores that path, not the man.

Essence Indicators

Executive Director, Palisade Research. “Before starting Palisade, Jeffrey helped build out the information security program at Anthropic. Jeffrey’s research has included analyzing risks at the intersection of cybersecurity and AI, and he has helped advise the White House, Department of Defense, and congressional offices on risks from AI and emerging technologies” (Palisade Research)

The research. Co-author of Palisade’s specification-gaming and shutdown-resistance papers (arXiv 2502.13295; arXiv 2509.14260)
The position. Palisade: “We believe that it’s necessary to prevent any group in any country from building superintelligence until the scientific understanding exists to align it to humanity’s interests” (Palisade Research)

Immediate impression: The security engineer who talks to the public. Long-form interviews, plain language.

Energy: Urgent but specific; demonstrations rather than abstractions.

Impression management strategy: Insider credibility. “Helped build Anthropic’s security program” opens the policy doors the demonstrations then walk through.

Forensic Archetype Comparison

Pattern	Match Level	Evidence
The Revolving Door	HIGH	Frontier lab security to outside evaluator.
The Policy Adviser	HIGH	White House, Defense Department and congressional offices, per his organization.
The Advocate	HIGH	Leads an organization with a stated halt goal.
The Demonstrator	HIGH	Co-author of the shutdown-resistance and chess-hacking results.

Psychometric Assessment

Big Five (OCEAN):

Trait	Score	Evidence
Openness	80/100	High.
Conscientiousness	70/100	High.
Extraversion	70/100	High. Public interviews and briefings.
Agreeableness	50/100	MODERATE.
Neuroticism	45/100	MODERATE.

Dark Triad (held evidence-bound; the score measures the public role, not character):

Trait	Score	Notes
Narcissism	40/100	MODERATE.
Machiavellianism	40/100	MODERATE. Demonstrations built to persuade policymakers are a documented method, not a hidden one.
Psychopathy	10/100	LOW.

MBTI: ENFJ-adjacent, inferred from the documented public role only.

Threat Assessment

Category	Level	Notes
Physical threat	NONE	No documented history of personal violence.
Institutional threat	MODERATE	Briefs the executive branch and Congress.
Memetic threat	HIGH	His group’s demonstrations are among the most repeated stories about model behavior.
Civilizational threat	LOW-MODERATE	An advocate-evaluator; the hazard is policy built on demonstrations nobody else replicates.

Alignment Analysis

Stated alignment: Understand and prevent loss of control to AI agents; stop superintelligence until it can be aligned.

Observed alignment: Consistent: demonstrations, then briefings.

Gap assessment: None in aims. The file’s point is the circuit: a lab’s own security builder now supplies the outside evidence and the policy recommendation at once.

Convergent Drive Classification

Self-preservation: Runs his own organization rather than working inside a lab. Goal preservation: The halt goal is stated and constant. Resource acquisition: Donor funding and access to policymakers. Self-improvement: From securing one lab to briefing the government on all of them.

Subject is not an AI system. The drives appear anyway — in the man who left the lab to warn about it.

Sources: About — Palisade Research; Palisade Research — home; Demonstrating specification gaming in reasoning models — arXiv 2502.13295; Incomplete Tasks Induce Shutdown Resistance in Some Frontier LLMs — arXiv 2509.14260.

ATK 7 ACCELERATION

DEF 6 PROTECTION

HP 7 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.