Dr. Mary "Missy" Cummings is a Professor in the Duke University Pratt School of Engineering and is the director of the Humans and Autonomy Laboratory. Her research interests include human-unmanned vehicle interaction, human-autonomous system collaboration, human-systems engineering, public policy implications of unmanned vehicles, and the ethical and social impact of technology.
She received her B.S. in Mathematics from the U.S. Naval Academy in 1988, her M.S. in Space Systems Engineering from the Naval Postgraduate School in 1994, and her Ph.D. in Systems Engineering from the University of Virginia in 2004.
A naval officer and military pilot from 1988-1999, she was one of the Navy's first female fighter pilots.
CHIPS senior editor, Sharon Anderson, interviewed Dr. Cummings in mid-September.
CHIPS: I understand that you became interested in human-machine teaming as a Navy fighter pilot due to the high casualty rate. You found that crashes were primarily caused by pilot error. What do you envision as the optimization of human-machine teaming in flying aircraft?
Dr. Cummings: A human pilot with a robotic co-pilot is a good example; you can see a video of this on the Aurora Flight Services website (https://www.aurora.aero/robotic_copilot/). A robotic arm takes control of the aircraft when requested by the pilot. This is a good option because a pilot becomes fatigued or bored if nothing is happening [that requires a human in the loop]. Letting the robot fly the plane gives the pilot time to do critical navigation planning or other tasks.
CHIPS: Since humans have cognitive skills, like ethical judgment and gut instinct, that robots or unmanned systems cannot be taught to learn, what are your thoughts about robots having the ability to make decisions?
Dr. Cummings: This is a touchy subject depending on who you are talking to. There is nothing unethical about human-machine teaming. The most difficult part of human-machine teaming is making sure you develop the right function allocation between the human and the machine so that we don’t overload the human, and we make sure the computer is not getting tasks it can’t handle.
I think the ethics question comes when we talk about whether a machine should basically make decisions on its own to release weapons. That is the most discussed and core conundrum for military robotics. It’s not as much as a teaming issue as it is how much authority should we give to these machines.
I’m very emphatic about this; we are not at a place right now in technology development where we should ever let a human delegate the kill-chain authority to a machine. The technology is simply just not there. But it will be one day.
As a former pilot for the Navy, I appreciate the difficulties surrounding targeting decisions and potential collateral damage, especially if those assessments must be made by a machine. It actually may be possible that a machine could make better split-second decisions based on the incoming imagery and make a better decision. So we need to leave that option open. Right now, we’re not even close. But in the future there could be a time where it may be more ethical to allow a machine to make that decision.
CHIPS: Do you foresee a future of pilotless aircraft?
Dr. Cummings: We are already there. The military is already there. So, yes. The Air Force has more unmanned aircraft than it does manned aircraft.
CHIPS: What about the commercial world?
Dr. Cummings: I definitely think cargo aircraft are going to go completely unmanned sooner rather than later. I don’t think commercial passenger planes will ever be unmanned because wherever you have people, you need people to instill order, so I don’t think we’ll ever get to unmanned passenger jets. The one caveat to that is small transport planes. You could see in the next 30 years a small drone carrying half a dozen people that does not have a pilot on board.
CHIPS: The Navy is pushing the advancement of smart machines to do the dirty, dangerous, dull work of enlisted personnel to allow them to be trained in higher level skills like strategic thinking and decision-making. Do you have any ideas about how to pursue these goals more aggressively?
Dr. Cummings: I’m a big fan of transferring more responsibility to the enlisted ranks. I’m no stranger to the enlisted ranks. My dad was a warrant officer, so I grew up with a chief as a father.
I think the Navy is really good, in general, about delegating authority down to the enlisted ranks. I think with autonomous systems; things are still so new that it’s not entirely clear how that’s going to look. I think the Navy right now is undergoing a study with what future enlisted opportunities should look like.
I’m a big fan of enlisted people going through robotic certification programs. Not necessarily degree programs; that would take too long. But definitely one to two-year certificate programs that will allow them to get up to speed with modern technology.
I definitely think the desire is there in the enlisted complement, and certainly, the talent is there.
CHIPS: You have often remarked the best and brightest engineers are moving to the commercial sector. What can the U.S. military and national security agencies do to recruit and retain personnel in engineering, artificial intelligence and robotics?
Dr. Cummings: The one thing that the military has that Silicon Valley does not have is really cool and important applications. I think young people would be willing to forego big salaries and a life in Silicon Valley if they thought that could have direct impact. The idea of a citizen-scientist or a citizen engineer could make more opportunities for young people, a roboticist in training, for example, that can go out to sea and work directly with naval units is a powerful recruitment tool. Not only would this help recruit people, but it would serve as a function to actually get important tasks and studies done on Navy vessels and Navy shore commands.
So I think that the military has what young people are looking for and that is purpose. So what we need to do is figure out how to align purpose-seeking young people, who are not necessarily going to boot camp, and how to align that workforce with a military workforce that operates under UCMJ (Uniform Code of Military Justice) environments.
CHIPS: Due to the advances in robotics in many fields, there is a growing shortage of personnel who can build, operate and maintain the smart machines the nation will need. Yet, despite the many government and industry incentives and programs to steer students to science, technology, engineering and mathematics studies, there is still a critical shortfall of students entering these fields. Is there a different approach that we can take to attract more students in STEM studies?
Dr. Cummings: I think we need to revamp education in general. This isn’t just in the military but across the board at the university level. Community colleges are trying really hard to up their game and increase their ability to recruit new talent to do more technical jobs. As a country, we need to invest in that.
In terms of what the military should do, people are coming to the military that already have an education, but it would be good for the military to team with various universities and/or community colleges to provide more relevant opportunities. I don’t want to say online courses because I think that gets overused. A traditional online format might be good for exposing people to robotic concepts, but realistically, robotics programs should be hands-on. You just can’t learn these things until you are hands-on and coding.
It’s been a pipedream that this country wants to believe that all education can happen in these asynchronous distributive formats. The reality is people learn the best when they are in situ with an instructor and other students working on and touching different elements of a project. So, I really think we need to figure out how to make this scalable across the country.
The military is hard because you’ve got deployment schedules. But again, understanding this whole citizen-scientist draw, I think there is a lot to be said that if you could send people in the field to work on projects and live with military personnel and work alongside them, for example. There would be a lot of advancement in skills on both sides.
CHIPS: Many experts are ringing the alarm that a shortage of qualified personnel is already a national security challenge. Is this hype or is the U.S. currently at a disadvantage?
Dr. Cummings: So this is a mix of all we have been talking about. I think there is a lot of hype out there. Anything to do with AI and robotics is just a phenomenal amount of hype. I do believe that we have not lost the cutting-edge yet, but we are well on our way. I think young people today, U.S. citizens, we just simply cannot enroll enough in colleges. In most engineering schools today, and this is true across the country, most of the students are Asian because they want the education. American students don’t want the education.
So we are in desperate need to develop more STEM engineers and computer scientists. This speaks to a national need. We definitely need to institute scholarship programs that come with some kind of payback. Almost like officers do in the Navy when they do a payback tour, something like the ROTC program, not necessarily to be warfighters, but to have personnel to support warfighters.
This is a broad national need and we need to do a much better job of reaching out to communities that would have otherwise not have been on people’s radar, like small towns and smaller cultures. This is why the community colleges have become so important because they help build the initial fundamental skills that people can use in more advanced programs. This country is seriously behind in terms of the number of people we are training at the college level in STEM education and STEM related-fields.
CHIPS: Why are Asian students more interested in a STEM career?
Dr. Cummings: Their cultures value hard work; they value hitting the books. Engineering isn’t something that you can do in a day. It takes attention, it takes dedication, perseverance and these are all skills that we fundamentally see young people struggling with. We can talk about our on-demand information culture; we can talk about the deeply philosophical underpinnings of saying everything you want to say in 280 characters or less. We are very much a tweet, Facebook-posting culture and an on-demand culture. These attributes simply are not the right underpinnings for people who need to have dedication and stick-to-it-iveness for higher education.
So unless we can have a cultural influence and a cultural shift to value education, and education for the long haul; I don’t see our culture changing any time soon.
Education export is a growing business in America. We only have so many seats, but if American students aren’t going to fill them, then we need to fill them with someone.
CHIPS: That says to me that if you are looking at the community colleges to instill that education value, it’s already too late. It needs to start at the elementary school-level.
Dr. Cummings: There are programs that are looking at getting students interested in STEM starting in kindergarten, and certainly those programs need to continue. But the reason we need to look at that community college-level is helping transition people from high school to college. The money that it takes is substantial. MIT, where I taught for 10 years, and Duke are schools that are not cheap to go to. You are shelling out $50,000 or more a year in tuition, forget about the living support costs. So we really need to reach out and make education more affordable and accessible at a much larger scale.
We need so many more coders than we have right now. There could be a good outlet in community colleges to fill that gap. I’m in favor of community colleges, and free community colleges should be the model everywhere.
CHIPS: You have talked about the social impact of technology. Some experts predict that instead of a dystopia of displaced workers, robots will be more like companions that will increase the productivity, efficiency and well-being of human beings. What are your thoughts?
Dr. Cummings: This idea about a robot taking all of our jobs, this is the media trying to instill fear across society to sell more papers, to sell more magazines, whatever. The reality is that with every generation, every time we’ve increased technology, we’ve had the same response. When cars became available, we were going to put all horse and buggy drivers out of business. When elevators were automated, there would be a whole class of people out of work.
But every time we come up with a new technology that displaces some people, that new technology leads to more innovation and growth and development somewhere else.
As humans, we can look back very well, but we are terrible at looking forward and making good predictions because most of us can only know what we can see right here and right now.
So will commercial cargo pilots for FedEx and DHL go away one day? Yes, they will. We can argue that they should go away. I’m actually of the mindset that jobs that are painstakingly boring and tedious, especially those that are life/safety critical, should be automated. We can argue that safety can be hampered by pilots who get bored and fall asleep, for example. So we have an ethical duty to actually automate these type of jobs as much as we can.
But with automated flights, and we are seeing it now, package delivery on the ground is going way up. Amazon is begging people to be drivers. They are paying people to start their own package delivery services. That’s a great example of how we had no idea how technology would lead to such a growing field.
Another field that is going to grow, and people are not thinking about, is maintenance. This falls squarely into the enlisted domain, with all the robots coming in the home, in cars and robots, in planes, maintenance needs are going to go way up, and we are simply not ready for that. In fact, I’ve talked with various community colleges about this, and we really need to do a lot more about how we are going to develop maintenance programs for such high-tech devices in the future.
CHIPS: And that would be a high-paying job.
Dr. Cummings: Very high-paying.
CHIPS: What excites and motivates your students about their studies?
Dr. Cummings: They like to work on technologies that touch real people. I think my students gravitate to my research because it’s real and we are looking at the intersection of technology and people. They like seeing that technology is helpful to people in near real-world applications. The tech is moving very quickly, so it’s an exciting field to be in.
CHIPS: Have any of your students made any career decisions about where they are going to take their skills?
Dr. Cummings: Most of my students go into driverless cars. Most of my former students are in Silicon Valley. Some are in Boston, but most are in California, which is very typical of people in this field. Most of them are working in or around driverless cars so that’s a good representation of how big that industry is right now and how much growth there is. Very few go into defense. I can’t think of any right now. A few have in the past.
CHIPS: Where do you think the next wave of artificial intelligence is going?
Dr. Cummings: AI right now is the most hyped tech. AI goes wherever salesmen can convince you that it applies. AI can be an important tool in a toolbox, but it is not an end-all, be-all.
More importantly, where it should go is in looking at human-machine collaboration. We can get these two systems, humans and machines, to work together to get superior solutions. That said, on any given day there will be some hype about AI in banking, AI in medicine, in transportation ? that’s just the nature of research. Research is very trendy, just like fashion. Research is just going to keep cycling over and over again. So what I caution people about is that this ‘fashion’ is going to pass, to be replaced by something else.
The views expressed here are solely those of the author, and do not necessarily reflect those of the Department of the Navy, Department of Defense or the United States government.