OpenAI taps former Army secretary Dan Driscoll for a national-security role

A former Army secretary joins the world's most powerful AI company

OpenAI has appointed Dan Driscoll, who served as the 26th Secretary of the US Army, to a newly created position titled senior counselor for national security modernization, the company said in a statement on Monday, November 2, 2026.

Driscoll's mandate spans the company's fast-growing national-security business: advising OpenAI's leadership, working with outside partners on how advanced AI can strengthen national security and cybersecurity, engaging government officials, industry leaders and academic institutions, and advancing what the company calls "democratic AI"; while also helping OpenAI's own teams understand how national-security organizations operate, Reuters reported.

"I'm joining OpenAI to help the United States and our allies put this technology to work responsibly, for the people who keep us safe," Driscoll said in the statement. The Singapore outlet CNA carried the Reuters report in full, noting the appointment extends the company's expansion into classified government work.

The hire lands two months after Driscoll's bruising exit from the Pentagon. OpenAI is no stranger to churn of its own; earlier high-profile safety resignations raised separate questions about the lab's direction.

The job description, in OpenAI's own terms

Five facts anchor this story. First, the title is new: "senior counselor for national security modernization" did not exist at OpenAI before Monday, which signals a dedicated seat for the company's government business rather than a portfolio folded into an existing executive's desk.

What does a "senior counselor for national security modernization" actually do?

Second, the duties are advisory, not operational: OpenAI describes Driscoll as counseling its leadership and collaborating with external partners, government officials, industry leaders and academic institutions, on practical initiatives for "democratic AI." Third, the February deal is the context that matters most: OpenAI agreed to deploy its models on the Pentagon's classified networks, with written bars against mass domestic surveillance, directing autonomous weapons, or making high-stakes automated decisions. Fourth, the company published separate principles for government partnerships in July. Fifth, Driscoll's résumé fits the brief unusually well: beyond running the Army, Trump tapped him for talks with Moscow and Kyiv; a rare diplomatic assignment for a service secretary, Reuters noted.

The appointment also arrives mid-race: labs, states and regulators are writing competing rulebooks for military AI on every continent.

What the announcement left unsaid

Reuters was explicit about the gaps: the company statement did not say when Driscoll starts, who he reports to, what he is paid, or which projects he will handle. That is a lot of blank space for an announcement whose headline is a person's name.

Nor is it clear whether Driscoll's remit is limited to the February agreement with the Pentagon or ranges more widely; into allied-government business, intelligence work, or the cybersecurity products OpenAI sells commercially. The July principles and the February exclusions set boundaries on paper, but who adjudicates edge cases inside the company has not been disclosed.

One more unknown matters: whether Driscoll's appointment required any ethics clearance or cooling-off review after his Pentagon departure, and whether Congress will ask to see it.

The Pentagon-AI revolving door just spun faster

Former officials joining AI labs is not new, but the seniority and the timing are. A sitting service secretary who left office in September joining the company with a classified-networks contract in November compresses the usual cooling-off narrative into weeks; and it hands OpenAI something money alone cannot buy: a man who personally managed the Army's bureaucracy and negotiated, however briefly, in the Russia-Ukraine file.

It also arrives as the Pentagon itself is reorganizing for AI-enabled warfare. The department's autonomous warfare command push shows where the institutional energy is going; and why the labs are competing to be the supplier of choice.

For the industry, the hire sets a benchmark. If OpenAI's bet pays off, every frontier lab will need a national-security counselor of its own; and every resigning Pentagon official becomes a recruitment target.

Reading the appointment through the five dimensions

Geopolitics: the move tightens the knot between American AI supremacy and American military power. Rivals will read "democratic AI" as a euphemism for allied-only advantage, and the appointment will sharpen the question Beijing and others ask about every US lab: how close is it to Washington's war planning?

Economics: classified AI work is becoming a market, not a mission: multi-year deployments on secured networks, compliance overhead, and premium talent salaries. OpenAI's February deal opened the revenue lane; Driscoll is the hire that staffs it.

Human capital: the Pentagon's turbulence, firings, forced retirements and resignations under Hegseth, is producing a supply of senior talent that the labs can absorb. The flow from government to industry, usually gradual, is accelerating.

History: the revolving door has precedents in aerospace and consulting, but rarely at this speed for a technology this young. What took decades in defense contracting is happening in the AI era within months.

Technology: models on classified networks are no longer experimental pilots; they are production infrastructure. The question is no longer whether AI enters the war room, but who writes the usage rules once it is there.

The ripples this hire will send

First ripple: recruitment pressure. Anthropic, xAI, Google DeepMind and the defense primes will all be asked why they lack an equivalent national-security counselor; and the answer will be another bidding round for ex-Pentagon talent.

Second ripple: oversight. Congress, which funds both the Pentagon and watches the AI labs warily, will want to know how the cooling-off rules apply when the departing official lands at a current contractor. Expect letters, then hearings, then rules.

Third ripple: the allies. Driscoll's quote mentioned "the United States and our allies"; allied governments will ask what access, if any, the counselor role extends to them, and on whose terms.

Fourth ripple: the conflict zone. Classified AI tools land in live operations; the US-Iran escalation showed how quickly military planning can move from memo to strike; which makes the February exclusions on weapons and surveillance the most important text nobody has tested yet.

What to watch in the weeks ahead

Watch for the start date and the org chart: who Driscoll reports to will reveal whether the role is a strategy perch or an operational command. Watch for his first public remarks, which will be parsed in three capitals.

When will Driscoll actually start?

Watch for copycat hires at rival labs within weeks, and for a congressional reaction; the timing, one day before the midterms, guarantees political scrutiny whatever the outcome at the ballot box.

Watch, above all, for the first public disclosure of what the classified-networks deal is actually producing: capabilities, guardrails, and the moment an exclusion is tested. That is when the "democratic AI" promise meets reality.

One final variable is the calendar. The announcement landed on the eve of the US midterm elections, when every personnel move in Washington is read as a political signal. Driscoll's arrival at OpenAI will therefore be interpreted through two contests at once; the battle for AI's military future, and the battle for Congress. Neither will stay quiet for long.