The Air Force is aiming for AI readiness by 2025 and becoming AI competitive by 2027, according to Chief Data and AI Officer Maj. Gen. John Olson.
To the department, being AI ready means having a trained workforce, enterprise IT and all the foundational training, infrastructure and readiness elements associated with it. Being AI competitive means technological superiority and granting combat forces strategic advantage on the battlefield.
Responsible, resilient and robust AI is essential to compete and win against global competitors and adversaries, Olson said Tuesday during Federal News Network's AI in Government panel.
"This is … the view all the way to the top, the deputy secretary of defense has AI and ML readiness and responsible AI as the number two defense priority," Olson said. "Our secretary of the Air Force and undersecretary of the Air Force are also absolutely focused on this as an enterprise, ubiquitous capability that we must have to succeed."
More than 600 AI projects were in the works at the Defense Department as of April 2021. DOD requested $874 million for AI development for fiscal year 2022, but that number doesn't reflect DOD's total AI investments. Lockheed Martin and IBM subsidiary Red Hat announced their collaboration to tackle challenges DOD faces in the AI space and advance AI innovation for the department yesterday.
The Air Force has several projects underway leveraging AI across a broad spectrum of use cases, including both predictive and preventative air maintenance, crew and aircraft scheduling, space imagery, data collection and human-machine teaming.
"We don't have just one significant AI or [machine-learning (ML)] program, we have a plethora of them and that's growing almost exponentially," Olson said. "We see data in AI/ML readiness and the operationalization of that is a critical, almost ubiquitous part of everything that we do not only now but as we evolve to the future."
One notable application of these capabilities within the Air Force will be the Advanced Battle Management System (ABMS), the department's contribution to DOD's Joint All Domain Command-and-Control (JADC2) initiative. As the Air Force looks at the convergence of these areas, JADC2 is "the ability to sense, make sense and act across all domains with the joint services in a contested environment at the speed of relevance," Olson said.
"As we look at and talk about wingman AI and collaborative combat aircraft — that's absolutely essential to what we're doing, but, of course, we don't even fly without maintenance and logistics and basing," he added. "These are really complex and challenging endeavors. And we're leveraging data and access to our authoritative data in analytics at speed with AI, ML capabilities. So these are really, really ubiquitous enablers."