A version of this story appeared in the daily Threat Status newsletter from The Washington Times. Click here to receive Threat Status delivered directly to your inbox each weekday.
The Pentagon’s departing artificial intelligence chief is offering a parting warning against tech companies’ claims that powerful new AI tools can serve as the solution to, or the cause of, existential problems.
Chief Digital and Artificial Intelligence Officer Craig Martell told a congressional hearing Friday before his April exit from the government of his rising frustration with companies pitching AI “magic.”
Mr. Martell told the House Armed Services Committee’s tech panel that developing AI will not necessarily save the country from its adversaries nor will America’s competitors’ obtaining the powerful new tool lead to unthinkable disaster.
“It’s extremely important to not see this as a monolithic technology, which is how it’s sold, ‘Get this thing: If we have it, we win and if they have it, we lose,’” Mr. Martell said at the hearing. “I think that’s fundamentally flawed. It is neither a panacea nor is it a Pandora’s box.”
Mr. Martell said AI claims should be evaluated by looking at particular uses for the tools, something he has sought to do in his nearly two years as the Department of Defense’s first AI chief.
A computer science expert with experience at the ride-sharing platform Lyft and file-sharing company Dropbox, Mr. Martell’s Pentagon perch gave him unprecedented visibility into cutting-edge AI tools under development by the country’s top tech minds. He said Friday he remains bullish on AI’s ultimate potential despite his own cautions.
He assembled leading AI companies, from Big Tech to smaller startups, in Washington last month to huddle with defense and intelligence officers, academics, and others. During the symposium’s unclassified portion, he acknowledged that the Pentagon needed AI companies’ help and he said the government should have reached out sooner to the private sector.
Asked by Rep. Mike Gallagher, Wisconsin Republican, on Friday about how to judge his tenure, Mr. Martell said he created a “massive increase in demand for getting it right.”
“A year ago, there were a lot of, ‘I don’t know what this means,’ ’I don’t know what you mean by getting the data right,’ ’I don’t know what you mean by having the data accessible to the warfighter at the right time in the right place,’” Mr. Martell said. “That’s now well understood.”
Within a month of Mr. Martell’s gathering of AI heavyweights in Washington, Secretary of Defense Lloyd Austin thanked him for his service and named his replacement, Radha Plumb, deputy under secretary of defense for acquisition and sustainment.
Ruffling feathers
Mr. Martell ruffled some feathers while in office. Last year, his office’s employees gave their leadership a negative rating, according to survey results obtained by DefenseScoop.
Mr. Martell told lawmakers on Friday that consolidating different department offices and making decisions about what to focus resources on antagonized some people and said his team emphasized execution over communication as it was getting established.
As his office works on transitioning to new leadership, he said his team is supporting the Pentagon’s Replicator project, which aims to put AI into weapons systems and the hands of military service members.
The department has a goal of fielding multiple thousands of all-domain autonomous systems via Replicator by August 2025. It is an ambitious objective but is meant to ensure the U.S. military is ahead in a world where Mr. Martell previously told The Washington Times that keeping a human in the loop is not always possible anymore.
Mr. Martell said his team was building a new testing environment for Replicator complete with data, scenarios and potential software tools.
“The folks who are part of the Replicator initiative will come and try out their software, their predictions, their detection on our data,” Mr. Martell said.
Mr. Martell’s office has partnered with the tech company Scale AI in the office’s efforts to help test and evaluate AI applications for U.S. troops.
• Ryan Lovelace can be reached at rlovelace@washingtontimes.com.
Please read our comment policy before commenting.