OpenAI hires former US Army secretary as military seeks to expand AI use

OpenAI has hired Dan Driscoll, the former US Army secretary who resigned following tensions with defence secretary Pete Hegseth, as the military seeks to expand its use of AI to enhance national and cyber security.

The Iraq war veteran and close friend of vice-president JD Vance will join the group in a newly created role as senior counsellor for national security modernisation, OpenAI told the FT.

The San-Francisco based group added that Driscoll would focus on working with American allies to advance the responsible use of AI.

“Dan will work with OpenAI teams and leaders across governments, industry, academia and allied institutions to advance practical modernisation initiatives to advance the build-out of democratic AI,” the company said.

It added that Driscoll would help its teams “understand the operational needs of global national security institutions”.

At the Pentagon, Driscoll drove reform efforts aimed at introducing new technology more quickly and making weapons more interoperable. Earlier this year, he organised a huge hackathon called Operation Jailbreak to improve communications between weapons using lessons learnt in Ukraine.

The 40-year-old former defence official who clashed with Hegseth over the Pentagon chief’s firing of several widely respected army officers has been widely tipped to play a senior role in any potential Vance 2028 presidential campaign.

“National security requirements are changing faster than at any time in the past 50 years,” Driscoll said. “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.”

OpenAI chief executive Sam Altman said the US had an “extraordinary opportunity” to work with countries to spread the adoption of democratic AI. “I’m excited to welcome Dan Driscoll to OpenAI to strengthen American leadership and help us do that.”

OpenAI said Driscoll would advise public and private sector partners in modernising policy and strategic ways to “unlock practical uses of AI to improve national security and help ensure the world is building AI to support democratic principles”.

His appointment comes as OpenAI has provided cyber defence in several European countries, including France, Germany and Poland, with access to its models. It has also allowed the Ukrainian government to access a programme called Daybreak designed to help identify software vulnerabilities.

The tools being developed by OpenAI and its rivals are regarded as critical to US national security, both as a form of defence and as a potential vulnerability.

Earlier this year, OpenAI expanded its relationship with the Pentagon, agreeing to deploy its AI models in classified settings following the collapse of negotiations between rival Anthropic and Hegseth.

Recommended

But the risks posed by AI have also become clearer in recent months, exposed by a series of high-profile hacking incidents involving OpenAI, including breaches of open-source platform Hugging Face and the Australian government.

OpenAI was widely expected to go public this year, having confidentially filed its IPO prospectus with the Securities and Exchange Commission in June. But it has deferred its listing until next year and is instead seeking to raise $20bn in a private funding round this month.

OpenAI cut its funding target from $30bn and the valuation it was seeking to raise at from $1.4tn to $1.25tn amid growing anxiety about the pace of AI investment.