Richard J. Joseph is Chief Scientist of the Department of the Air Force, advising the Secretary and Chiefs of the Air Force and Space Force on their $2.8 billion science and technology enterprise. He directed the development of the Air Force’s 2030 Technology Strategy for former Air Force Secretary Heather Wilson, completed in 2019. On Jan. 28, he spoke with Air Force Magazine Editorial Director John A. Tirpak about the structure of the S&T enterprise, emerging game-changing technologies, and the future of humans in the loop. The conversation has been edited for length and clarity.
Q. You argued that the Air Force should create a Chief Technology Officer position, but that hasn’t happened. Why not?
A. We took that cue from conversations with industry, and it’s also just good sense. You want someone who is deeply in touch with technology development but is also in touch with the customer; with the product centers. And in our case that would be the Majcoms and the COCOMs, and also industry.
And we said, this has to be somebody who controls the science and technology program, and maybe the RDT&E program, and it needs to be somebody who reports to the Secretary, somebody at a high level, like an assistant secretary, and there needs to be a counterpart on the Air Staff.
There was a lot of opposition to it. And the acquisition office was the strongest opponent of that because they felt it was still important to have that “birth to death” oversight of systems development.
I understand that, but I didn’t think—we didn’t think—it would work. … And we think the laboratory system could produce better work if it was more focused on the research and development, science and technology area, rather than contract management of things that were supposed to become programs of record.
Q. How about for the Space Force?
A. In the early formation of the Space Force, I asked Gen. (John W.) Raymond, actually before he was designated for his current job [chief of space operations], have you thought about a chief technology officer? And he said, no, what would that person be? And so I explained.
Months later, he had created this chief technology and innovation officer on his staff, and I congratulated him on that. He said, ‘the more I thought about it, the more sense I realized it made.’
That wasn’t my original idea, that was people on our executive review panel for the 2030 Technology Strategy.
Q. How would you break up the S&T responsibilities to work better?
A. (Gen.) Jimmy Doolittle and Gen. (Hoyt) Vandenberg got the idea that doing research under the direction of people who buy systems wasn’t the best way to do it; that the acquisition system tends to be ‘now’ focused: what can we do now, what can we work on now, what can we start now. Vandenberg and Doolittle felt that was too confining, that we would not really be able to harvest the gains in technologies made in both the commercial and academic area.
That was changed in about 1950. They created an assistant secretary for research and development, and a deputy chief of staff for development. And what followed was almost three decades of really strong innovation. In the 1990s, the Air Force decided to consolidate research under the acquisition system. Since then, we’ve had less aggressive science and technology research and exploitation of new ideas, and acquisitions have taken longer and longer and been more expensive. And I think those are related.
This is just me speaking, but I would …[suggest] having breathing room between the people who do programs and [those that] do…program management … of contracts with industry.
There are three real important things the [Air Force Research Laboratory] does.
No. 1, they address short-term needs that the MAJCOMs come up with, and some mid-and long-term needs. And those are real important and serious. Most of that, by the way, is done by managing contracts and industry.
The second one is to do the research. Somebody needs to keep us on the cutting edge of science and technology. That is typically done by the AFRL staff as well. However, the people who are managing these contracts with industry … are the same people who are supposed to be doing the research.
Now, everybody has collateral duties. But if you are managing $20 million a year in contracts and you’re doing part of a $5 million a year research effort, you know where the emphasis is going to go.
So I would advocate splitting those a little bit. Meaning, we set up a program management operation, and it can have people in it who have been in the research and development area for a long time; sort of like the DARPA model. They are running these programs, they’re not actually trying to do research themselves, but they are knowledgeable about the research because they have strong backgrounds in it.
Some are doing research and some may cross over and become program managers. And we could build a sort of firewall—a soft firewall—between the program management people and the research and development people. They talk to each other, they use each other, but we don’t have researchers who are doing program management. And by the way, most of them really don’t like doing that, so it’s a retention and a recruiting issue. That’s my prescription.
Q. We’ve been talking about hypersonics, directed energy, smallsats, etc., for 40 years. Those things seem to be close to fruition. What will be the defining Air Force technologies 40 years from now?
A. Well, I hope that in 40 years we’ll have those things you mentioned, taken even further.
I think the big things that are coming, is materials by design. Which means we’ll have tools that allow us to formulate and fabricate materials, their structure and composition, in ways that we could never do before. I think that’s going to substantially change things.
The last 50 years have been the decades of physics; the next 50 years will be the decades of biology. We need to really exploit that. Synthetic biology will result in new materials, new understanding of the human, and it will hopefully also address a new understanding of our cognitive processes.
For instance, who owns your cognitive profile? The way you think, the way you see things? What you see when you look at a scene, or when you read something?
When you talk to a used car salesman, he’ll ask you a few questions and develop a profile of you based on how you answer the questions. If someone can screen your emails and phone conversations and speeches and the articles you write, and they develop a pretty detailed understanding of how you think, do you own it? Or do they own it? Do they get to do what they want with it because you freely offered up the way you think?
That’s the cognitive side. The cognitive side will also benefit things like artificial intelligence (AI). For example, when a pilot puts their CAC (Common Access Card) into the slot, it will identify who they are but will also identify their cognitive profile. Things they see well, and things they do not see well. And then the system adapts to that pilot. The next day, it could be a different airplane, a different pilot, and it adapts to that person.
And, in a broader sense, we will have better ways of predicting societal actions, including … leadership decisions, and maybe even political decisions.
Q. What about propulsion, things that fly, things that fly in space? Is hypersonics the end of the line? And are we “there yet” in directed energy?
A. Well, I don’t think hypersonics is quite hyper enough. To really have some differences we will continue to push that envelope: higher and higher speeds, as well as the ability to maneuver. This is contingent on new materials, and our ability to model the dynamics of those systems. Such as, how do the forces come into play when they change the direction of a missile? I think that’s going to continue to develop. There are many more development cycles to come.
Directed energy is really at a point where we can make use of it, and not just for weapons, but also for sensing. We’ve been driving so hard to get very high power levels out of compact lasers, we sort of missed that there are a lot of things we can do in really small packages that are really useful. Sort of like—but not—the LiDARs (light detecting and ranging) that are in cars today, and are being talked about for autonomous vehicles. But maybe much more sophisticated than that.
Q. For example?
A. The LiDAR that is used for collision avoidance in your automobile—which can be used as a target designator—really just calculates how far away things are, by knowing the transit time of the signal going out and coming back.
But lasers can also interact with a material and tell you something about what it’s made of. We’ve been trying to exploit that for a variety of things over the years. One was for spotting clouds of biological organisms and identifying them from unique signatures. Very hard to do; they were relatively heavy if we wanted to do it at long ranges, and it was hard to put them in a helicopter or in an airplane.
But then it shrunk, and it will continue to shrink. And as it does, we will exploit other signatures. So, lasers will be big in the sensing area.
The same with microwaves. We already have more information than just the distance and location from radars, but we may push that dimension even harder.
Q. Not long ago, the Pentagon said it was shifting its top priority from hypersonics to microelectronics. What’s the long view there?
A. Most of our advancements in AI and machine learning have come about because the microelectronics community did their job superbly. They stayed on the path of Moore’s Law for decades. And there were times when it looked like they had hit a serious dead end, but they’ve somehow always come through with a way to get by it. They deserve a big part of the credit for any successes that AI is having.
We will continue to research microelectronics. We will go way beyond, I believe, some of the approaches that we’ve taken to date; neuromorphic computing just being one of them. And that’s where the biology comes in.
Q. Are we going to have a jack in the back of our skull where we’ll plug into our computer or airplane?
A. God, I hope not. I’m not going to do that.
AI is going to be important. But it is an enabler. It’s not a panacea, there’s no magic. But it will get better and better.
Q. The Air Force is already using AI on airplanes, for mission planning and execution. We’re on the cusp of the robotic “loyal wingman.” Are we coming to the end of the period where we have human beings in the airplanes? Are people just going to provide guidance for the machines from a rear location?
A. I’ve thought a lot about that, and I don’t think so. It’s not because I don’t think AI or a lot of those things will become really capable. It’s because all those advantages we talked about—hypersonics, electronic warfare, etc.—they all have a big machine learning and network component to them.
But networks are vulnerable. Computer systems are vulnerable, and what happens if we have a conflict where these advantages are challenged? What if we lose some of our advantages through cyber [attack], and through kinetic kills in the air on the ground and in space? If this happens on both sides, who has the advantage then?
If you have people in the system, then you have a pretty good computer in there. And if you’ve had experience and you’ve been trained well, then you aren’t bereft of all of your advantages.
So, I don’t think manned systems are going away completely. I think there’s still an important future … maybe even a more important future for manned systems, so pilots can rest easy. But not too easy. I know we can do wondrous things and that will augment our capabilities, but it also creates vulnerabilities. I think manned systems are with us for a while.