DARPA, the US Defense Department’s research arm, will spend $2 billion over the next five years on military AI projects.
Is that a lot? That depends. In the realm of AI research, it’s a huge chunk of change—China, for example, made waves earlier this year when it announced it was putting a similar amount of money behind an AI-focused research park in Beijing. The Pentagon, though, functions on a different scale—$2 billion isn’t quite enough to buy 20 of its shiny new F-35 fighter jets.
The focus: The initiative, announced Friday, is being billed as a way to form better partnerships between machines and humans. DARPA said new projects will focus on things like security clearance vetting, improving the reliability of AI systems, and exploring explainable AI. Notably absent from the press release: any mention of autonomous weapons.
The subtext: The new funding comes in the wake of strong Silicon Valley pushback against working on military AI programs (see “Google won’t renew its military AI contract”). Some researchers are taking the announcement as DARPA saying, in effect, that if Google won’t accept its money, that’s no problem. It’ll find someone else that will.
This article first appeared in The Download, our daily tech newsletter. Sign up here.