OpenAI hires former Army Secretary Dan Driscoll to lead its AI and national security strategy

OpenAI has hired Dan Driscoll, former secretary of the US Army, as its first senior advisor for national security modernization. He arrives two months after leaving the Pentagon and with a very specific assignment: to put artificial intelligence at the heart of allied defense and cybersecurity.

The contract with the Pentagon that explains the signing

It is not a cosmetic gesture. OpenAI already sells to the War Department (the name that the Trump administration has recovered for the Pentagon) the integration of ChatGPT into GenAI.mil, its corporate artificial intelligence platform. The tool is available to millions of military and civilian employees for operational planning, mission preparation and administrative tasks.

Their Driscoll position is newly created. The mandate is to examine where AI improves operational needs, how it drives institutional change, and how much it shortens the production of results. In parallel, it will accompany the allies of Washington in the so-called ‘responsible use’ of technology.

The message of Sam AltmanCEO of the company, was explicit: ‘AI must give people more power over their own lives.’ The stated objective is to export the American technological stack: ‘The United States has an extraordinary opportunity to work with other countries to expand the adoption of democratic AI.’

The former secretary formulated it in military terms: ‘The people who serve deserve the best tools we can put in their hands. And he added the fact that explains the rush: ‘National security requirements are changing faster than at any time in the last fifty years.’

His time in the Army was marked by modernization and an obsession: autonomous systems. He earned the nickname ‘the drone man’ from donald trump. He resigned on August 31, and the published versions point to differences with the vision of the Secretary of War, Pete Hegseth; the president appointed Adam Telle as acting secretary. The signing closes the circle between the Army and Silicon Valley.

The movement does not measure the power of a technology: it measures who is already writing the regulations for the next war.

What Spain is at stake in this race

Spain, for the moment, watches from the sidelines.

The reason is structural. Europe has legislated a lot and bought little: The European AI Regulation set the rules and Spain created its own supervision agency, the AESIA, based in A Coruña. But rules do not create models. When an allied army wants an assistant to plan a mission, today it buys in California, not Amsterdam.

For the Spanish defense industry, the movement redefines the terrain. Indra, Telephone, Navantia, ITP Aero either Hispasat They compete as integrators for contracts that are negotiated within the framework of the European Defense Fund: sensors, communications, cybersecurity, maintenance. The founding model almost always arrives importedand with it the conditions of use.

The calendar is tight. Spain committed to reaching 2% of GDP in defense spending in 2029 and 5% in 2035, and that money will have to be distributed between physical platforms and software. The question that is asked Madrid It is not whether there will be military AI, but who will supply it and with what access to the code.

The European comparison should be made without embellishment. France has Mistral; Germany pushes its own ecosystem with public money; Spain has talent and an early regulator, but no border model with which to sit at the table of a General Staff. That changes the map.

Washington Logic

Understanding the signing requires looking at how defense buys USA since 1945. Unlike France or Spain, where the State maintains industrial champions, Washington chose a private contractor model: the Pentagon defines the requirement and the company provides the technology. The move of a former Secretary of the Army to a technology company does not break that tradition: it confirms it. Eisenhower warned of the military-industrial complex in 1961; McNamara went from presiding over Ford to running the Pentagon. The revolving door predates Silicon Valley.

The underlying logic is speed. The problem the Pentagon is trying to solve is not technological, it is bureaucratic: buy software with procedures designed to buy aircraft carriers. With Driscoll on board, the company positions itself on the side of the table where military requirements are translated into commercial product. And it does so with an argument that sells in this city: ‘democratic AI’ as an export instrument versus China.

Behind it there is a broader coalition than it seems. Republicans see efficiency and deterrence; a part of Democratic Party requires civilian oversight; Export control hawks see a way for allies to buy American batteries instead of Chinese ones. Everyone is pushing in the same direction.

There are risks, and they are real. Relying on a single supplier in a critical area is a national security vulnerability. Driscoll’s time in the Army was brief and his departure was surrounded by conflicting versions. And the concentration of federal AI contracts in a handful of companies —OpenAI, Anthropic, Google, xAI— converts any conflict of interest into a state matter. As CSIS remembers, absorbing commercial technology today outweighs the number of troops.

For Spain, the reading is different. We are not facing a threat, but rather a race of standards that can be lost by default if Brussels continues regulating without buying and Madrid continues buying without requiring technological transfer. The useful window is short: the next European budget cycle and the review of the defense programs that are negotiated throughout 2027. If the Spanish Army ends up using assistants manufactured in California, the data will be decided there.

The projection is verifiable. The Pentagon’s fiscal year 2028 budget request will arrive at Capitol at the beginning of 2027, and there it will be seen if Congress sustains the pace of commercial AI purchases. The calendar rules more than the speeches.

Case File

• The case: OpenAI brings in former Army Secretary Dan Driscoll as senior national security modernization advisor, two months after his departure from the Pentagon.
• Key data: newly created position focused on defense, cybersecurity and cooperation with allies; the company already integrates ChatGPT into GenAI.mil for millions of employees; Driscoll resigned on August 31.
• For Spain: The national industry competes as an integrator in European defense programs, without its own foundational model with which to negotiate access conditions.