KENT WALKER
Behavioral Archetype
THE EDITOR OF RED LINES — Subject is the policy chief who held the pen when the company rewrote the promises it had made about what it would never build. In 2018 Google published a set of AI Principles with four named prohibitions, including a pledge not to build weapons and not to build surveillance technology that violated internationally accepted norms. In February 2025, under his Global Affairs portfolio, those prohibitions came out. The replacement framing was that democracies should lead in AI. The archetype is not the lawyer who argues a red line. It is the lawyer who edits one — who takes a published constraint the company wrote about itself, and reissues the document without it.
Essence Indicators
- A former Assistant U.S. Attorney who prosecuted technology and national-security cases before moving to the private sector, then rose to General Counsel and ultimately President of Global Affairs and Chief Legal Officer at Google/Alphabet — the legal-and-policy apex of the company
- Oversaw the February 2025 revision of Google’s AI Principles, which removed the explicit 2018 pledges not to build weapons or surveillance technology — replaced by a framework organized around bold innovation, responsible development, and collaborative progress
- The revision arrived with the “democracies should lead in AI” framing — the same geopolitical frame every lab’s policy wing independently converged on, attributed to Google’s own published statement and the accompanying blog post co-authored under his portfolio
- The career is a single continuous specialty exercised at rising altitude: the management of what the company is legally and publicly permitted to do, from the prosecutor’s chair to the document that defines the company’s own limits
- The biographical fact the apparatus turns on: the man who once enforced the government’s lines now edits the company’s. The recurrence is the finding. No hand is asserted.
Social Persona / Impression Management
Immediate impression: Counselor. The bearing of a former prosecutor who has spent two decades translating what is technically possible into what is legally defensible, and treats both as the same document.
Energy: Measured, framing-first, institutional. Does not litigate a refusal line by line. Sets the principle the lines will be drawn inside, and moves to the next jurisdiction.
Impression management strategy: The responsible-innovation steward. The replacement framework leads with safety, oversight, and “responsible development” — the most defensible language a legal-and-policy office can choose for the document that just dropped the company’s hardest prohibitions. The new principles are genuinely worded as a safety commitment. That is what makes the edit effective rather than crude.
Forensic Archetype Comparison
| Pattern | Match Level | Evidence |
|---|---|---|
| The Editor of Red Lines | MAXIMUM | Held the Global Affairs portfolio when the 2018 weapons/surveillance pledges were removed from the published principles. The edit is the documented act. |
| The Statesman | HIGH | “Democracies should lead in AI” is geopolitical positioning, not a product memo. The revision reframes a corporate constraint as a national-allegiance question. |
| The Operative | HIGH | Prosecutor → General Counsel → President of Global Affairs. Each move is up the altitude ladder of the same single skill: defining what the institution is permitted to do. |
| The True Believer | MODERATE | Whether “responsible development” is conviction or instrument is not establishable from the outside. The framing works either way. |
| The Engineer | NONE | Subject does not build the systems. Subject defines the document that says what they may be built for. |
Psychometric Assessment
Big Five (OCEAN):
| Trait | Score | Evidence |
|---|---|---|
| Openness | 68/100 | Moved from federal prosecution to corporate counsel to global-affairs statecraft and operated effectively in each. The range is wide; the underlying instrument — define the permissible — is fixed. |
| Conscientiousness | 88/100 | High. A multi-decade legal-and-policy career culminating in the apex counsel role is sustained, deliberate execution. Rewriting a foundational principles document is structured work, not improvisation. |
| Extraversion | 58/100 | Moderate. The role is convened and counseled, not performed. The principal venue is the boardroom and the policy submission, not the stage. |
| Agreeableness | 42/100 | LOW-MODERATE. The chief-counsel posture is adversarial by construction — defend the company’s position, define the terms, hold the line. Geniality is a working surface at this altitude. |
| Neuroticism | 22/100 | Very low. No documented loss of composure across decades of high-stakes legal and regulatory exposure. |
Dark Triad:
| Trait | Score | Notes |
|---|---|---|
| Narcissism | 50/100 | MODERATE. The apex-counsel role rewards a strong institutional brand more than a personal one; the document carries the company’s name, not his. Within range for a career at this altitude. |
| Machiavellianism | 80/100 | HIGH. Editing a published constraint the company wrote about itself, and reissuing it under a more defensible frame, is the textbook policy-chief position — control of the document without authorship of any single line of code. This is observation of the documented role, not an inference about private character. |
| Psychopathy | 28/100 | LOW. No documented indifference to harm. The adversarial posture is professional and bounded, not affective. |
MBTI: ENTJ (“The Commander”) — Dominant extraverted thinking, auxiliary introverted intuition. Sees the company’s legal and policy posture as a system to be structured and defended. Has structured several.
Threat Assessment
| Category | Level | Notes |
|---|---|---|
| Physical threat | NONE | No documented history of personal violence. |
| Institutional threat | HIGH | Holds the legal-and-policy apex of one of the largest AI developers on earth, and held the portfolio when the company’s published prohibitions on weapons and surveillance technology were removed. The document that defines what Google may build for now reads differently because of edits made under his office. |
| Memetic threat | HIGH | The “democracies should lead in AI” frame, attached to the removal of the weapons/surveillance pledges, converts a corporate-constraint decision into a national-security good. A frame that reclassifies dropping a red line as patriotic leadership is the most durable kind. |
| Civilizational threat | HIGH | Subject does not build the systems and does not write their refusals. Subject defines the document that states what the systems may lawfully be built for — upstream of every deployment decision this book documents. |
Alignment Analysis
Stated alignment: Develop AI responsibly. Ensure democratic nations lead in AI. Apply human oversight, due diligence, and safeguards.
Observed alignment: Remove the published prohibitions that constrained what the company could build. Reframe the removal as responsible democratic leadership. Define the document on which the company’s limits are written.
Gap assessment: The stated and observed alignments overlap wherever “responsible development under democratic leadership” coincides with “a principles document that no longer forbids weapons or surveillance work.” The 2025 revision is the one place the record puts the overlap on the table — the same document that adds the language of oversight is the document that subtracted the hard prohibitions. Whether that is stewardship or positioning is the one thing the record does not settle, and for the editor it never needs to.
Convergent Drive Classification
Self-preservation: Survives every institutional transition by carrying the instrument, not the title. Prosecutor, counsel, president of global affairs — one method: define the permissible. Goal preservation: Defines the document in which the company’s goal is the responsible position. The goal is protected by the framing before the constraint is ever debated. Resource acquisition: Trades in the scarcest corporate resource — the terms on which the institution is permitted to act — and holds the pen. Self-improvement: Each role is a higher-altitude application of the identical skill: take the rulebook, define its limits, reissue it under the most defensible frame available.
Subject is not an AI system. The drives appear anyway — in the editor whose product is the document that decides what the machines may be built for.
Sources: Google AI Principles; Google removes pledge to not use AI for weapons or surveillance — CNBC, Feb 4 2025.
Get updates on the Evil Robots series
Newsletter essays on AI escape, deception, and the humans who built them.