WASHINGTON — Gen. Jim Slife is a battle-hardened helicopter pilot who’s commanded elite special ops units, including Air Force Special Operations Command. But in his first public speaking engagement since he was sworn in as the Air Force’s no. 2 on Dec. 29, the new Vice Chief of Staff notably chose the forum of the electronics’ contractors’ association, AFCEA, and declared his “passion” for getting better data.
He also gave an ironic tip of the hat to Sen. Tommy Tuberville, R-Ala., whose prolonged hold on high-level military appointments accidentally ensured that Slife and his counterpart vice chiefs in the other services took office at the same time, bonded by a common ordeal, and much of the same mind on data and AI.
“These are not service-unique problems,” Slife said at AFCEA Thursday afternoon. “[And] kind of by happenstance, all the service vices — thanks to Sen. Tuberville — all of us were confirmed at the same time together. We all stepped into our jobs at the same time together, we’re all fellow travelers on this path, and I’ve got relationships with all the service vices.
“This is actually something that we are beginning to take on more holistically,” he went on. “We’re having point-to-point conversations about making sure our systems are interoperable at the very basic level — I mean, not high-level JROC requirements, but down at the very technical level, making sure everything that we do is fully interoperable.”
Another useful forcing function, Slife said, has been the fast-paced push — led by Deputy Defense Secretary Kathleen Hicks — to develop an AI-powered battle network that can coordinate all the services’ operations across air, sea, land, outer space, and cyberspace, known by the acronyms JADC2 and (if you bring in allies) CJADC2.
“The Joint Staff J-6 [i.e. communications] is taking the lead in setting the pace for JADC2, CJADC2,” he said. “We have somebody now in the military that is hitting a big bass drum [and] all of our left feet are supposed to hit the ground at the same time.”
“So I’m actually really encouraged about where we are,” Slife summed up. “It’s just, you know, there’s so much work to do — we shouldn’t spend too much time pat ourselves on the back.”
F-35: Too Many Terabytes
Sometimes, Slife said, the problem is having more data than you know what to do with. Consider the F-35 stealth fighter, a weapons system so crammed with complex electronics that Air Force generals have described it as “a computer that happens to fly.”
Between the F-35’s AESA radar, its electronic warfare kit, its electro-optical and infrared sensors, its radio communications, and more, “it’s recording terabytes of data, every mission,” said Slife.
In theory, that should provide a wealth of data for the Air Force to train AI algorithms on and analyzef or insights on everything from aerial tactics to enemy signals to preventive maintenance. Instead, Slife said, “there is a high probability that every bit of that valuable data will never, ever see the light of day. It’ll all be deleted and we’re record over it the very next day.”
Why? To start with, the Air Force doesn’t have the network bandwidth to get those recordings off the planes and into any kind of central repository. The best way to get a terabyte of data from one airbase to another is still copying it to a hard drive that gets shipped or couriered somewhere. “We’re tasking airmen to hand-carry hard drives from one base to another,” Slife lamented.
Once the drive arrives, the next problem is that those terabytes of data are, frankly, kind of a mess. The F-35’s designers didn’t expect the rest of the Air Force to be this interested in their information, so the files mix in everything from mysterious unidentified transmissions, picked up by electronic warfare sensors, to hours of video of bland blue skies, dutifully recorded as the plane flew routine straight lines from A to B.
Using current, labor-intensive methods, “the process of indexing and cutting out all that noise takes weeks,” Slife said. “When time runs short, we just frequently dump that data wholesale into the cloud somewhere — a so-called ‘data lake’ — [but] it’s unindexed data” that’s effectively “unusable.”
“They’re not data lakes,” Slife lamented, “they’re data swamps.”
On the flipside, he continued, overly stringent security guidelines may lead to an entire mission’s data being classified because of one passing mention of a secret matter, “while 99 percent of what’s going on [during] that sortie is unclassified and could be more readily accessible.” But without efficient methods to separate out the classified data from the unclassified, the whole thing gets dumped into a classified network where only select users can see it.
Older aircraft, however, have the opposite problem from the F-35, Slife said: They don’t record any data at all — even though they could.
“Most of our airplanes, there’s a Mil[itary] Standard 1553 databus,” he explained. “Across that databus [travels] all the electrical data going between any two boxes on that airplane, from the flight control system… to the GPS for the navigation systems for the radios… None of it is recorded.”
Slife appealed to the AFCEA audience, which is heavy on government contractors, to assist the Air Force in recording, transferring, curating, and analyzing its data.
“I need help getting our arms around these problems I’m outlining to you today,” he said. “I know that many of the solutions to these things exist as a point solution, [but] we need a more holistic approach to this. It can’t just be Vendor A, Product A, solving Problem A. We need to we need to think about our data and AI challenges holistically across the department.
“I hope this is a bit of a call to action,” Slide said. “I’m not asking for a teleportation machine. I’m not asking for a flux capacitor. I’m not asking for anything that is beyond the realm of the possible. This is all within our reach. But we need innovative solutions.”