DEAN BALL

Behavioral Archetype

THE PRIVATE GOVERNOR WHO WENT INSIDE — Subject is the policy writer who argued, from think-tank desks and a weekly newsletter, that frontier AI should be governed first by private bodies overseen by government, who then spent four months drafting federal AI strategy at the White House, who returned to a think tank to argue for broad preemption of state AI laws and for licensed outside auditors of the frontier labs, and who on 6 July 2026 became the head of a policy-strategy team inside one of those labs. His own announcement gives the reason: “In order to advance those intuitions, to make them more concrete, I realized that I simply must “go inside.”” (Hyperdimensional, 18 June 2026). This profile scores the position, not the person.

The position is the seam between the author of an audit proposal and the audited. His proposal says “it would be good for someone to thoroughly audit the frontier labs at least to test their adherence to their own safety plans” (Hyperdimensional, 26 June 2026). He wrote it ten days before starting at a lab, and labelled it “my own and not as an official position of any industry actor, including my future (but still not current) employer, OpenAI.” He is also a fellow at Fathom, the sponsor of California’s evaluator-designation statute, a non-resident senior fellow at the Foundation for American Innovation, whose Vice President of Growth sits on the board of AVERI, and a co-author of AVERI’s flagship auditing paper. Every one of those ties is on a page he or the other party published. None is evidence of what any designation, standard or bill will do, and he discloses most of them himself.

Essence Indicators

• Head of Strategic Futures, OpenAI. His announcement: “on July 6, I will be joining OpenAI at Head of Strategic Futures, a new team created to shape frontier AI policy within the company.” The team is “a small, high-agency team within OpenAI, reporting to Chief Strategy Officer Jason Kwon, charged with shaping frontier AI policy, which is to say matters pertaining to: catastrophic risk, recursive self-improvement, labor market impact, and the relationship between the frontier labs, governments (particularly the U.S. Federal Government), and society” (Hyperdimensional, 18 June 2026). His newsletter’s About page, now: “I am the Head of Strategic Futures at OpenAI and a Nonresident Senior Fellow at the Foundation for American Innovation” (Hyperdimensional, About). His FAI profile gives the same (FAI, Dean Ball). The team’s blog launched 20 August 2026, with the note that “the words below only reflect my views and not the views of OpenAI or any other of its employees” (Hyperdimensional, 20 August 2026).
• Non-Resident Senior Fellow, Foundation for American Innovation. FAI’s own announcement of the OpenAI move, signed by Zach Graves and Max Bodach: “Dean will stay on with FAI as a non-resident senior fellow, so this is closer to an expansion of our relationship than an exit.” FAI credits “his writing at Hyperdimensional, to his co-authorship of the White House AI Action Plan, to the body of AI governance and technology research he built here at FAI” (FAI, 18 June 2026). See Foundation for American Innovation.
• Author, Hyperdimensional. Launched 9 January 2024: “Hyperdimensional is a newsletter about emerging technology, public policy, and the future of governance” (Hyperdimensional, 9 January 2024). His stated perspective: “I write from a classical liberal and un-apologetically pro-progress perspective” (Hyperdimensional, About). He told readers of the OpenAI move: “the writing I do here will be fully mine: no one at OpenAI will have preapproval or editorial discretion over what I write here,” that “I can publicly depart from the positions of OpenAI and its leadership,” and that “As an OpenAI employee, there may be legal barriers to my commenting publicly on such a topic” as litigation involving the company (Hyperdimensional, 18 June 2026). In January 2026 he opened a paid tier: “This subscription is for firms seeking private, one-on-one or small-group conversations with me about the various matters of AI and AI policy I cover in this newsletter,” at “$7,500 per year” (Hyperdimensional, 15 January 2026).
• Former Senior Policy Advisor for AI and Emerging Technology, White House Office of Science and Technology Policy. FAI’s biography: “He most recently served as senior fellow at FAI, senior policy advisor for Artificial Intelligence and Emerging Technology at the White House Office of Science and Technology Policy, and strategic advisor for AI at the National Science Foundation” (FAI, Dean Ball). MeriTalk, as reported on 11 August 2025: “Ball, who joined OSTP in mid-April, said he is rejoining the Foundation for American Innovation (FAI),” quoting his post of that day, “The AI Action Plan is out, and with that I will be returning to the private sector” (MeriTalk, 11 August 2025). His own account: “I also pointed them out while I was a staffer in the White House, and wrote about them as part of my contributions to this administration’s official AI strategy, the AI Action Plan” (Hyperdimensional, 26 June 2026).
• Fellow, Fathom. On 27 January 2025: “I am pleased to inform you that I will be taking on a consulting role with Fathom, a new nonprofit working on AI governance and organizer of the recent (and excellent) Ashby Workshops, as a fellow,” with the note that “this new affiliation does not affect my role as a fellow at either the Mercatus Center or the Foundation for American Innovation” (Hyperdimensional, 27 January 2025). In June 2026: “Fathom, an organization with which I have an affiliation” (Hyperdimensional, 26 June 2026). FAI’s biography lists “a policy fellow at Fathom” among his earlier positions. See Fathom.
• Co-author, AVERI’s flagship paper. The author list of “Frontier AI Auditing: Toward Rigorous Third-Party Assessment of Safety and Security Practices at Leading AI Companies,” dated January 2026, names Dean W. Ball with the affiliation Foundation for American Innovation, beside Bri Treece (Fathom) and Grace Werner (AVERI). The paper states: “inclusion as an author does not entail endorsement of all claims in the paper, nor does authorship imply an endorsement on the part of any individual’s organization” (AVERI, Frontier AI Auditing). AVERI’s team page lists no advisory or staff role for him (AVERI, Team).
• Earlier posts and other seats. FAI’s biography: “a research fellow in the Artificial Intelligence & Progress Project at George Mason University’s Mercatus Center and a policy fellow at Fathom”; earlier “senior positions at Stanford’s Hoover Institution, the Calvin Coolidge Presidential Foundation, and the Manhattan Institute, where he oversaw the Hayek Book Prize”; “Dean serves on the board of directors of the Alexander Hamilton Institute and was selected as an Aspen Ideas Fellow” (FAI, Dean Ball). His forthcoming book: “my forthcoming book, to be published next year with Penguin Press” (Hyperdimensional, 18 June 2026).

On the Record

His words, from his newsletter and his employer-neutral posts. Positions argued as his own.

On governing by private bodies, 27 January 2025. “That, for now, America’s approach to AI should rely just as much, if not more, on private governance rather than laws, rules, and regulations promulgated by governments.” “Private governance of AI does not mean industry governance of AI.” “we do not currently know how to govern AI, and we should generally avoid using government and public policy to accomplish objectives we cannot define.” And the cost, stated by him: “To some extent, it requires us to trust the AI field—the industry, its investors, academia, non-profits, and others—to play a meaningful role in governing itself” (Hyperdimensional, On Private Governance). His safeguard: “it will be important to establish clear transparency rules to ensure that the field can be held accountable by the public, as well as clear oversight by government of private governance institutions” (same essay).

On auditing the frontier labs, 26 June 2026. “So, it would be good for someone to thoroughly audit the frontier labs at least to test their adherence to their own safety plans, as well as to probe their broader internal governance of the AI recursive self-improvement loops.” On who should do it: “A private body not tied to the changing political valence and foreign policies of the American federal government is likelier to be able to earn that kind of buy-in than an auditing function run by the government itself.” On the state’s role: “Perhaps, if you are persuaded that private bodies have merit, you could imagine the government certifying or licensing the auditors, just as auditors in accounting are licensed.” On the lever: “Governments could require that the largest AI labs receive certifications from these bodies in order to participate in the AI field. Alternatively, government could use a carrot approach: a lab with a certification could enjoy a safe harbor in liability contexts, for example.” On the name: “Fathom, an organization with which I have an affiliation calls them independent verification organizations.” On the weak point, in his own words: “aren’t these private bodies ultimately just verifying a lab’s compliance with its own safety and security framework? And this is a legitimate criticism.” On the end state: “Perhaps eventually the private bodies could write the “canonical” safety standards, or perhaps it could be a government agency. On this I am currently neutral” (Hyperdimensional, What Should Be Done). He names the existing field: “the ecosystem has already been emerging for years in the form of organizations like the Frontier Model Forum, AVERI, METR, Apollo Research, Fathom, the AI Underwriting Company, and others,” and the House text: “Something very close to this was just proposed by Representatives Obernolte and Trahan in their Great American AI Act.” See METR, Frontier Model Forum and AIUC.

On the source of the idea, 16 June 2026. “This notion dovetails with the work of Gillian Hadfield, whose writing on regulatory markets (both in general and with respect to AI in particular) is the primary inspiration for my own work on private governance.” And: “Key to Hadfield’s idea is that these private bodies would exist in competition with one another, allowing regulation itself to evolve with technology, societal attitudes, and the like.” On the design goal: “One hope of mine is that there are ways to design institutions that minimize political interference in technocratic matters, which has been the focus of my writing on private governance and independent verification organizations.” And a limit: “Now, I am not saying we need an “FDA for AI,”” (Hyperdimensional, Leviathan Waking). See Gillian Hadfield.

On preemption, 2 October 2025 and 15 January 2026. His proposal: “Establish a three-year learning period, during which no state laws targeting AI in any of the following ways may be enacted: algorithmic pricing, algorithmic discrimination, disclosure mandates, or mental health.” Paired with a duty on the developers: “Create a transparency requirement on frontier AI developers, as measured by their annual spending on AI-related research and development, to disclose their approach to frontier safety evaluation and risk mitigation.” And a limit on the law’s reach: “It does not regulate the development of the technology itself and instead requires developers to tell the public how they are building the technology” (Hyperdimensional, “Be It Enacted”). On liability: “We don’t need new laws for “accountability,” a euphemism in policy circles for “liability.” We’ve already got them.” His ask: “my proposal asks our political class to exercise, for only a few years, that scarcest of civic virtues: restraint.” In January 2026: “I have long been a supporter of a federal AI law with broad (though not universal) preemption of state regulations,” and “preemption really must be expansive, if short of an outright ban on state AI regulation” (Hyperdimensional, The AI Patchwork Emerges).

On himself and the lab, 18 June 2026. “some of the most important frontier AI governance decisions are likely to be shaped by, if not fully made by, the frontier labs themselves.” On the register of the role: “at all points in my career, I have tried to match the tone and tenor of my public communication with my role” (Hyperdimensional, That Untravell’d World).

Social Persona / Impression Management

Immediate impression: The classical-liberal institutionalist. His About page presents a newsletter by a writer who is “skeptical of an expansive government role in the economy” and who believes that “government’s primary functions are to ensure the security of its citizens, uphold the rule of law, and protect property rights” (Hyperdimensional, About). The biography lists essays on Oakeshott and Confucius and a Hamilton board seat (FAI, Dean Ball).

Energy: Essayist with a policy-drafter’s output. FAI credits him with a House Foreign Affairs Committee testimony on semiconductor-equipment export controls, comments to the General Services Administration on a proposed AI-safeguarding clause, and a role in an open letter on nucleic-acid screening signed by several lab chief executives (FAI, Q4 2025 update; FAI, Q1 2026 update; FAI, Q2 2026 update). Cadence since the OpenAI role: “2-3 posts per month, though perhaps higher if circumstances allow” (Hyperdimensional, 18 June 2026).

Impression management strategy: Disclosure by announcement. Each move arrives with a post that names the affiliation, states the independence terms and names the limits. The Fathom role was announced in the essay that introduced private governance; the OpenAI role in a post that says he cannot comment on litigation involving the company; and the proposal for licensed auditors carries the line “Please treat this proposal as my own and not as an official position of any industry actor” (Hyperdimensional, 26 June 2026). The Fathom announcement is in the 27 January 2025 essay. The strategy is transparent in form. What it cannot do is test itself: the promise that “no one at OpenAI will have preapproval or editorial discretion over what I write here” is a statement, not a record (Hyperdimensional, 18 June 2026).

Forensic Archetype Comparison

Psychometric Assessment

Trait scores are observations of the documented public role, not claims about private character. The public record is a weekly-to-monthly newsletter, a think-tank biography, one co-authored paper and FAI’s accounts of his work.

Big Five (OCEAN) (inferred from the public role only):

Dark Triad (evidence-bound; the score measures structural position, not character):

MBTI: INTJ-adjacent, inferred from the documented public role only. The rationale is the work product: a system-builder who writes long-form arguments with a numbered proposal at the end. His About page states the method: “My goal is for this newsletter to be useful to you in understanding the technological change happening all around us” (Hyperdimensional, About). A style inference from his writing, not a test result.

Threat Assessment

Score move: none. Calibrated against Gillian Hadfield (6/7/7, hazard 66.7), the author of the regulatory-markets design he names as his “primary inspiration”: level on leverage, two points more insulated on a named university chair and a listed grant base, one point more durable on a book, two papers and a chaptered statute; against Helen Toner (6/8/8, hazard 80), level on leverage, three points more insulated on a philanthropic general-support base and a centre she directs, two more durable on that centre; against Bri Treece (5/5/6, hazard 53.3), Fathom’s President, level on insulation and durability and one point below on leverage because she holds no drafting credit on a federal strategy; and against Jack Clark (8/8/8, hazard 82), who runs a lab’s policy organization and testifies on it, two points more leverage, three more insulation, two more durability.

Score basis. Atk 6 is leverage: he leads a team “charged with shaping frontier AI policy” across “public-facing policy (for example, proposals for legislation) and internal governance within the lab,” and FAI credits him with “co-authorship of the White House AI Action Plan”; the points under a higher number are that no document gives him a vote, a gate or a designation power, the preemption text and the audit design are proposals, and the House text is a discussion draft (Hyperdimensional, 18 June 2026; FAI, 18 June 2026). Def 5 is insulation: his seat is an employee’s role at the company whose policy it shapes, his FAI fellowship sits at an organization whose own page lists OpenAI among its corporate supporters, and his protection is his stated term that “no one at OpenAI will have preapproval or editorial discretion over what I write here”; the points above 3 are a paid newsletter tier, a Penguin Press contract and two fellowships that give him income outside one employer (Hyperdimensional, 18 June 2026; FAI, About). Hp 6 is durability: the AI Action Plan FAI says he co-authored, a posted preemption text, a newsletter archive since 9 January 2024, a co-authored audit paper and a book due from Penguin Press; the points off are that the Action Plan is an executive document, the House text is a discussion draft and the audit design is Hadfield’s, and Fathom’s, not his (FAI, 18 June 2026; AVERI, Frontier AI Auditing; Hyperdimensional, Leviathan Waking).

Alignment Analysis

Stated alignment: AI should be governed first by private bodies, in competition, overseen by government, so that rules can be found by experiment, with state law preempted for a learning period and developers made to disclose how they build; the state should not run the audits itself, and should license those who do (Hyperdimensional, On Private Governance; Hyperdimensional, What Should Be Done; Hyperdimensional, “Be It Enacted”). Observed alignment: He has put those designs into essays, a posted bill text and a House-draft credit, taken consulting and fellowship seats at the organizations that carry them (Fathom, FAI, AVERI as co-author), served in the White House office that wrote the national strategy, and now works for a lab inside the class an audit would cover (FAI, Q2 2026 update; Hyperdimensional, 18 June 2026). Gap assessment: None on direction. The proposals are the same across 2025 and 2026, the Hadfield debt is stated, the Fathom tie is stated and the editorial terms are stated. The record raises a structural question and does not settle it: an audit proposal’s author is employed by a company that the audit would cover and that his own page lists as a supporter of the think tank where he holds a fellowship, while his stated terms give his employer no editorial control over what he writes. The pages show disclosed seats and nothing more, and for a file that scores a position it never needs to.

Interest: Documented, never motive, never a verdict. He argues three positions.

1. Position: licensed private auditors of the frontier labs. (“So, it would be good for someone to thoroughly audit the frontier labs,” Hyperdimensional, What Should Be Done.) What adoption pays: the California statute creates a designated class, “an AI auditor that is designated by the agency as having demonstrated expertise in assessing the risks posed by an AI system or model” (leginfo, SB 813); Fathom, the sponsor, is “an organization with which I have an affiliation,” in his words (Hyperdimensional, What Should Be Done), and he took a consulting fellowship there, whose compensation, if any, is not on any page read; AVERI, which lists Fathom among its funders, published the paper he co-authored, and his FAI colleague Grace Werner is on its board (AVERI, About; AVERI, Team); his own proposal names a carrot for certified labs, “a safe harbor in liability contexts,” and the lab that employs him states support for “SB 813 on overall infrastructure for independent safety assessments” (OpenAI, AI policy window). No fee, designation, contract or equity to him from any of them is documented. His best case: “The groups that do this will need to be composed of extreme technical experts,” and a government agency cannot pay or move like that; “A private body not tied to the changing political valence and foreign policies of the American federal government” can earn international buy-in (Hyperdimensional, What Should Be Done). Others who gain: the evaluators he names, the Frontier Model Forum, AVERI, METR, Apollo Research, Fathom and the AI Underwriting Company; any lab that would hold a certification and its safe harbor; and the legislators who need not write technical tests into statute.
2. Position: a federal AI law with broad preemption of state law, for a learning period, paired with developer transparency. (“I have long been a supporter of a federal AI law with broad (though not universal) preemption of state regulations,” Hyperdimensional, The AI Patchwork Emerges.) What adoption pays: a three-year bar on state AI laws on pricing, discrimination, disclosure mandates and mental health would lift those obligations from every developer subject to them; FAI’s own page lists corporate supporters and event sponsors including “Alphabet, Amazon, Anthropic,” Meta, Microsoft and OpenAI, the last his employer (FAI, About); his text in exchange requires frontier developers “to disclose their approach to frontier safety evaluation and risk mitigation” (Hyperdimensional, “Be It Enacted”). No payment to him is documented. His best case: “The rapaciousness with which states have sought to regulate AI is unprecedented for an emerging macroinvention,” and “the technology is too nascent for novel statute”; the bargain is “as modest a cost to bear in exchange for three years of state preemption as one could possibly hope” (Hyperdimensional, “Be It Enacted”). He also states the limit on his own case: “preemption is generally best suited for domains where there is clear consensus on the optimal regulatory framework. There is no such consensus here.” Others who gain: every developer that now faces differing state rules; the House authors of the discussion draft FAI says is “structured like” his text; and the White House, whose December 2025 executive order on state law he says faces “firm limitations on what the executive branch can unilaterally do” (Hyperdimensional, The AI Patchwork Emerges).
3. Position: voluntary pre-deployment testing, not a mandatory “FDA for AI.” (“Now, I am not saying we need an “FDA for AI,”” Hyperdimensional, Leviathan Waking.) What adoption pays: FAI’s quarterly update describes his position as “the case Dean has made that the federal government should evaluate frontier models for catastrophic cyber and biological capabilities through voluntary pre-deployment agreements, rather than a mandatory “FDA for AI” licensing regime,” and records that CAISI has signed evaluation agreements with Google DeepMind, Microsoft and xAI, “extending its voluntary pre-deployment reach to all five leading U.S. frontier labs” (FAI, Q2 2026 update). A voluntary regime leaves release timing with the developer. He notes the cost of the alternative in a June 2026 post: the executive order “which claimed to establish a voluntary testing program for frontier AI models, was really establishing a de facto involuntary licensing/preapproval regime,” and “OpenAI’s GPT 5.6 is being limited to only a small set of US companies at the request of the US government” (Hyperdimensional, What Should Be Done). His best case: “One major problem with this, as implemented, is that nobody knows what the requirements are to get licensed.” Others who gain: every developer that would otherwise face pre-clearance; the national-security agencies he says should keep testing for threats where “the most relevant knowledge resides principally within the government.”

Documented critiques, in their authors’ words. The record contains no critique addressed to him by name that this file found. The written critiques are of the licensed-auditor design in committee analyses of California’s SB 813. A coalition of the California Initiative for Technology & Democracy, the Children’s Advocacy Institute and Tech Equity on the 2025 version: “It is not clear whether there are any entities currently in the United States other than perhaps the sponsor of this legislation who might be able to comply with the current MRO qualifications in the measure” (Senate Judiciary analysis, 29 April 2025). TechNet, same analysis: “Rather than establishing bespoke organizations and frameworks from scratch, we encourage California to build on the work of established, globally respected standards-setting bodies” (Senate Judiciary analysis, 29 April 2025). On the independence premise, Fathom’s own July 2026 piece quotes Rumman Chowdhury: “Being a purely independent evaluator is nearly impossible, because it’s just not a fully built-out ecosystem yet” (Fathom, Who Evaluates the Evaluators?). His own essay names the weak point: “aren’t these private bodies ultimately just verifying a lab’s compliance with its own safety and security framework? And this is a legitimate criticism.”

Alliance Map

The overlap is the finding. At one time the same person is a lab’s head of policy strategy, a fellow of the organization that sponsored the California designation statute, a fellow of a think tank whose Vice President of Growth sits on the board of the auditing institute he wrote with, a co-author of that institute’s paper, the former staff author of the White House strategy whose director sat on his think tank’s board, and the author of a preemption text that a House discussion draft is said to follow. The documents show disclosed seats, listings and fellowships, not coordination, and no designation has been made. His principal intellectual debt, in his own words, is to Gillian Hadfield, who advises the same sponsor.

Convergent Drive Classification

Self-preservation: Holds concurrent seats across a lab, a think tank, a governance nonprofit and a publication, so no single loss removes his platform; each seat is stated in his own posts (Hyperdimensional, 18 June 2026). Goal preservation: Keeps one governance design moving across venues: a 2025 essay on private governance, a 2025 preemption bill, a 2026 audit proposal and a 2026 lab team charged with the same questions (Hyperdimensional, On Private Governance; Hyperdimensional, “Be It Enacted”). Resource acquisition: A paid institutional tier, a Penguin Press book, fellowships, and now a lab salary, none of which has an amount on any page read (Hyperdimensional, The AI Patchwork Emerges). Self-improvement: From a think-tank fellow to the White House AI office to the head of a lab team charged with policy and internal governance, by his own account to “go inside” (FAI, Dean Ball; Hyperdimensional, 18 June 2026).

Subject is not an AI system. The drives appear anyway — in the draftsman who went inside.

Public footprint: Hyperdimensional · FAI profile · X, @deanwball (linked from his FAI profile).

Sources: Hyperdimensional — About; Hyperdimensional — Archive; Hyperdimensional — Welcome; Hyperdimensional — On Private Governance; Hyperdimensional — “Be It Enacted”; Hyperdimensional — The AI Patchwork Emerges; Hyperdimensional — Leviathan Waking; Hyperdimensional — That Untravell’d World; Hyperdimensional — What Should Be Done; Hyperdimensional — “These Parchment Barriers”; FAI — Dean Ball; FAI — Dean Ball Joins OpenAI; FAI — About; FAI — Q4 2025 update; FAI — Q1 2026 update; FAI — Q2 2026 update; AVERI — Frontier AI Auditing; AVERI — About; AVERI — Team; Fathom — Who Evaluates the Evaluators?; MeriTalk — Dean Ball Departs OSTP; ProPublica — The Foundation For American Innovation; screendna.org — open letter; leginfo — SB 813 text; Senate Judiciary Committee analysis of SB 813, 29 April 2025; OpenAI — AI policy window.