OpenAI has appointed former U.S. Army Secretary Dan Driscoll to advise on national security modernization as it expands classified government work, Reuters reported Monday, citing a company statement.
Reuters said the newly created post will cover advice on how advanced AI could be used in national security and cybersecurity. Driscoll is expected to advise OpenAI leaders, work with government, industry and academic partners, and help company staff understand national security organizations' operational needs. Reuters did not report when he will start, who he will report to, his compensation or which projects he will handle.
The role adds a former Army leader to OpenAI's government work. Reuters did not identify a specific deployment tied to the position or say whether Driscoll's remit is limited to the company's agreement with the U.S. Department of War.
In February, OpenAI said it had reached an agreement to deploy its models on the department's classified networks. The company said the agreement bars use for mass domestic surveillance, directing autonomous weapons or making high-stakes automated decisions.
OpenAI's July principles set out a separate framework for government partnerships. The company said those partnerships should reinforce democratic accountability, meaningful human judgment and the rule of law. The principles describe its stated approach; the February agreement lists specific limits on classified-network use.
In the statement cited by Reuters, Driscoll said he joined to help the United States and its allies “put this technology to work responsibly” for people who keep them safe.
Driscoll was sworn in as the Army's 26th secretary on Feb. 25, 2025, after Senate confirmation, according to his official biography. The Associated Press reported that Sept. 2, 2026, was his last full day in the post, after about 18 months. The AP said a public post from Driscoll gave no reason for his departure.
