31st Annual Automotive Insights Symposium | How Automation Might Impact Auto Industry Competitiveness

KRISTIN DZICZEK: So our next moderator and the panel are going to come to the stage here. Kristen Broady and I started on the same day almost three years ago. Two Kristen's in the onboarding process together. We have actually three Kristens in research. So I'm Regional Kristin. This is Economic Mobility Kristen. And then back at the office, we have Micro Kristin, not because she's particularly small but because she's in microeconomics. So we can see her. She's not microscopic.

So Kristen is a Senior Economist, Economic Advisor, and Director of the Economic Mobility Project. And as Rick said, she has a substantial body of work on the impacts of automation on industries and workforce. And I'm going to go sit down and listen to what she has to say. Thank you, Kristen.
 
KRISTEN BROADY: Thank you, Kristin.
 
KRISTIN DZICZEK: [LAUGHS]
 
KRISTEN BROADY: So overall at the Chicago Fed, I believe there are five or six Kristens and five or six Dans. I don't know what it is, really common. But thank you, Kristin. Thank you, Rick. I also always have to thank Lauren and Peter in the back because they are how we all registered and got here. So I always want to recognize them. So I'm going to be talking with two really brilliant engineers about how automation might impact auto industry competitiveness. So short, interesting story. We have about an hour.

So I was sitting at my desk and I get this email from Carl and Ed, two people who I don't know. And they are reaching out to me-- it's like March of 2023-- about my research on automation and labor. And I get these emails all the time about interviews or whatever. So they're going to be doing something with robotics, their arms that reach out. And there are these hands, these end effectors, dexterity, and haptics, and a bunch of other words that I, as an economist, am looking up because I don't know.

And so fast forward to September of 2023. We're doing a site visit because the NSF is interested in this. Oh my goodness, right? Like the NSF does not fund many of these and it's a lot of money. Who knows how much? I don't know. And so August of 2024, I jump up from my desk and go to my boss-- who is the number one Dan in my book, my favorite boss ever-- to tell him that Ed's grant was funded for $26 million, like $26 million from the NSF.

And Dan was like, wait, what? The hand thing? The haptics thing? And I said, yes, $26 million. And he was like, well, what are you going to be doing? And I was like I don't know. I'm going to wait to hear from Ed. And so one of the things that I'm doing is this. So we had a kickoff in November. And so I am very pleased to introduce Ed Colgate and Carl Moore, two of my favorite engineers, partially because I don't know many, and, two, because they're amazing.

[LAUGHTER]
 
So I'm going to ask them to come forward as I introduce them. So Dr. Edward Colgate is the Walter P Murphy Professor of Mechanical Engineering and Director of the NSF Engineering Research Center on Human AugmentatioN via Dexterity, HAND, at Northwestern University. His principal research interest is human-robot interaction. He has worked extensively in the areas of haptic or touch interface, remote manipulation, and advanced prosthetics.

He, along with collaborator, Michael Peshkin, is the inventor of a broad class of collaborative robots known as cobots. So he's currently working in the following areas-- tactile, texture, recording, and reproduction-- which he's going to tell you about. I cannot-- high-speed electroadhesive clutches for applications, and haptics, and robotics, and robotic manipulation with dexterous hands. He holds a PhD in Mechanical Engineering from the Massachusetts Institute of Technology.

Dr. Carl Moore is a highly-respected mechanical engineer who has worked in the field with Fortune 500 companies like Kodak and Ford. Dr. Moore is currently shaping the next generation of engineers and is an Associate Professor of Mechanical Engineering at the Florida A&M University and Florida State University College of Engineering. Dr. Moore earned his Master's and PhD in Mechanical Engineering at Northwestern University. He is also a member of the HU7, a group of seven students that earned their engineering degrees at Howard and later earned PhDs.

So I am going to add a little bit about labor and how what they do impacts labor. But I'm going to mainly ask a bunch of questions of them so that we can learn what they do. All right.

AUDIENCE: And use the portal?

KRISTEN BROADY: Yes. And please use the portal for questions. I'm glad I did that. I'm only 5' 1". All right. Here we go. So I am very excited to be talking to you all. So I've talked some about your research interests and I mentioned the HAND Research Center. But can you tell us more about your research? And I'm not exactly sure what all of the backgrounds of the people here are. Some of them maybe make robots. Some of them may not make robots for a living. So can you tell us what it is that you do? And, Ed, I'm going to start with you.
 
ED COLGATE: Sure. Thank you, Kristen, for inviting us to be part of this. I feel a little bit out of place. Been listening to some just brilliant speakers today. I think the only thing I'll have in common with Austan was where we received our PhDs, so I'm thankful to have some connection there. But, yeah, I am an engineer. I'm a mechanical engineer who works on robotics.

And through my career, really all of which has been at Northwestern, I've really focused on human-robot interaction, particularly physical interaction, direct interaction between people and robots. And that work really has its origins in work that we did starting in the mid-90s with the automotive industry. My colleague, Michael Peshkin-- who you mentioned-- and I were approached by colleagues from General Motors, and then eventually, Ford, to tackle the following problem.

At that time, the automotive industry had very successfully used robots to automate welding up of a car body and to automate painting of a car body. But they had tried and failed spectacularly to automate final assembly, putting all the parts into the car on the assembly line. That was primarily done manually, still is today primarily. That's changing now. But and at any given day at GM-- this is mid-90s-- they told us that 19,000 workers were at home.
 
Some of them had the flu. A lot of them had a hurt back or shoulder, some sort of cumulative trauma. And so they came to us and said, hey, can we use robots to somehow help people, to improve the ergonomics, the working conditions? And that was the problem that we tackled. And it led to the invention of this class of machines that we called cobots. And today, cobots is a pretty big segment of the industry. We can talk about that evolution.

But we actually started a business in the late '90s called Cobotics where I spent a lot of time in this city and all around the city working with both the OEMs and the supply chain to look at how we could bring the benefits of robots-- and think about what they are, I mean, robots can be strong, they can be tireless, they can sync up to data in your factory in ways that may be inaccessible to people-- but bring those benefits together with the benefits of people, which, of course, are many as well. And so that's really kind of what got me started on a journey.

Along the way-- and I might want to come back to this later on-- I had the chance to join the board of directors of a publicly traded company in the automotive supply chain. So listening to Austan's discussion of supply chain just now resonated with me in many ways. I spent a decade working with that business and seeing how the supply chain was dramatically reshaped in a global fashion.

And then fast forward even more. In recent years, I've teamed up with Carl and a number of other colleagues to establish this new NSF-funded center. We call it HAND. That's an acronym of sorts. It stands for Human AugmentatioN via Dexterity. And what we're really after is a continuation of what began with that work on cobots, which is how do we bring robots and people closer together and really get the best of both worlds?

And in particular, the thinking here is that robotics today is a little bit like what computer mainframes were in the 1970s. I mean, they're pretty good computing machines in the 1970s, but they're expensive. They're accessible to really only the largest businesses who could afford the programmers, who could write the custom code, who had the scale to take advantage of it.

And it was really the advent of the personal computer that changed things. And it wasn't just making a computer smaller. It was really making it different. It was providing applications that worked, that provided functionality right out of the box-- get your computer, turn it on, and you can use it-- and, of course, also the user interface that made it accessible to the non-programmer.

And so we are asking ourselves in the center, how do you take a robot, in particular, a robot manipulator-- something that moves things around-- and make it accessible to not just the Fords, the GM's, the Stellantis, but to the companies throughout their supply chain so that they can buy that robotic arm, get it out of the box, and instantly, be up to speed and making use of it. That's a pretty tall challenge, but that's the revolution that we think is feasible in today's age. And so we'll be happy to talk more about that.
 
KRISTEN BROADY: Yeah, I guess as I think about that and listen to that, getting this arm and hand out of the box, I think about it being able to know the difference between a carton of eggs-- now that they're $10 I don't how many will be--

[LAUGHTER]

--versus a can of Coke. That just seems really cool to me. When you described it, that's kind of what came to my mind. So, Carl, I'm going to come to you. What are you working on?
 
CARL MOORE: Well, Kristen, I just wanted to say first, thank you for inviting me. It's been great so far. Economics is something I've always been fascinated by. Even though I was a robotics student at Northwestern, I feel like economics was always what everybody was talking about. And so it's very good to be here. It wasn't easy to be here. When they opened up the registration and told me to go ahead and register as a speaker, every time I went to register, it would say it's full. And that happened like three times, to the point to where finally, you just registered for me. So I appreciate that.

Like Ed said, I started my robotics journey in college. And one of the things that's always fascinated me about robotics is that you could bring a lot of your thinking to it. The things that you do as a human being every day, you were tasked with trying to find out what was the best way to translate that to a machine. And so you start off with some knowledge-- hey, I do this all the time, so I could be able to teach this machine how to do it.

And one thing that we learned very early on is that we could cover a lot more ground if we allowed the human and the robot to work together. And at that time, a lot of people were really focused on full automation, so turning on a robot and having it just go and do some stuff. But that exempted the human who was already so knowledgeable about whatever the process was.

And one of the things I first remember at Northwestern when we were working in cobotics on these collaborative robots was that there was-- I can't remember if it was Ford or GM-- but there was this assembly where they were bringing in dashboards. But as they brought it in, there was fire-retardant material that had often flapped down. And when the human saw that, he would just push it back up and go about the business. But it was so hard to get the sensing right to get that to be 100% automated.

And then that also brought into some of the things that Ed mentioned with people having repetitive-motion injuries. And so the cobot was able to basically work with the human, take some of that weight of this big item that the person was having to insert. So the cobot's holding the weight. The person didn't have to worry about that anymore. And the person just guides it into place. And if that fire retardant material folds down, the person just moves it out of the way. So it was a first time that I saw that human and robot come together in what we now call human augmentation, instead of automation, be spectacular and effective.

KRISTEN BROADY: So I'm not coming to this from Ford or GM or any other automotive company. I originally wanted to be a state trooper. Law school ended up not working out, so I went to work for Walmart. And I was a loss prevention manager at a really, really big distribution center in New Albany, Mississippi. And so loss meant a lot of things. It meant workers' comp, it meant destroyed products. And so as the LP manager for this whole place, I had to learn just about every job.

And so the one thing that I think I'll always remember is being an order filler. So I had to learn how to drive a PE. I didn't do the forklift training. But if you don't what a PE is, you're riding on something that moves pallets. You could put two pallets on there. And so you've got the RFID thing. And again, this is like 2002 2003. I'm older than I look. All right. So anyway, you've got your headset on and it tells you go to P44, huge pallets, pick two. You don't what it is. It doesn't matter. Two cans of Coke, pick it, put it on your pallet. Then go here and pick that.
 
Now, the problem is you're driving this huge piece of equipment with all of these pallets and you're just now learning, right? So when you learn how to drive, if you hit a cone, do you hurt yourself? No. If you run into something-- because at the front of this thing, there's no motor in front-- the first thing that is going to come into contact with something if you hit it is your hand. Oh God, guys, I'm driving this thing, and I run into the corner of one of the shelf systems.

We had not had a workers' comp case in more than three years, right? So managers had to carry around a can of spam. If you got caught without the can of spam, you would have to pay $1 because we were raising money for United Way, and also this had something to do with the workers' comp thing-- so raising money, and no workers' comp, I cannot be the one. I'm new here. And so when you reached out talking about a hand, I don't think I ever told you this story.
ED COLGATE: No you didn't.

[LAUGHTER]

 
KRISTEN BROADY: But I'm like if there was something that could hold onto that PE, or a joystick, or something such that your literal hand was not going to get crushed, or so that when you're picking something up, all of the back injuries from bending over and lifting or bringing something down and hoping it doesn't fall on you, or when something does fall because things do fall, anyway. When you called, I was like I remember this 20 years ago.
 
ED COLGATE: Now I understand, yeah.
 
KRISTEN BROADY:
 
And if this keeps people from having to do that or to work with anhydrous ammonia which is how a lot of places are cooled, or having to go in the how-many-ever-degrees below zero, I was like, I'm on board.

The second part of my research is looking at how stuff like that is going to impact workers. And so people often ask me-- so there are these automation risk scores that Frey and Osborne created-- so how many workers are going to be displaced in the next two years, three years, five years? And I'm always like I don't know. But what you're talking about is not necessarily displacing workers like myself or those order fillers, but helping them. And I think that's something that's really important.

So in that same vein, what impact are these things going to have on workers? We've sort of talked about it. And I'm sure in the automotive industry, there's stuff like order fillers and people who do those jobs. So how are these things going to work with workers? And what, if any, jobs might be displaced?
 
ED COLGATE: Yeah, that's a good question, and, of course, a very difficult one to answer. There's conflicting evidence on how bringing in automation technology, and specifically robots, does impact jobs. There's a very, very strong tendency, and I think if you follow the popular news, especially if you follow Elon, it's hard to overlook this these days.

There's this growing tendency to think of robots as directly replacing people in the sorts of work that we do. And so I highlight Elon because Tesla is developing a humanoid robot. And you'll see lots and lots of articles written about this. And it's very clear what the idea is. We will use automation to straight up replace people.
 
But if you actually look carefully at the history of, of automation and particular robotics, it's not often the case that things work out quite like that. Usually, the change is much more holistic. Changes happen in a way an overall manufacturing plant is set up, for instance, that creates new jobs, new opportunities. If you look especially at the smaller businesses in the supply chain that we're very interested in, and you look at who's actually adopting robots, and almost always it's businesses that are growing, that are thriving.
You look at those who have never used robots and automation, usually these are firms that have fairly stagnant opportunities and are not thriving as much. So I'm not trying to be a Pollyanna and paint this glowing picture of automation because it's very hard to tell. If Elon's vision is correct and humanoid robots are soon among us, well, why would they not take our jobs, all sorts of jobs? Why would they not? And we can get into this. I don't happen to think that's likely to be the outcome.

But anyway, so more to the point right now, I think the question about jobs that's more interesting really is one of supply chains. As we're looking at, we're talking a lot about tariffs, talking about supply chains, talking about manufacturers, as this whole thing gets reshaped, and we try to reshore manufacturing, and we try to create opportunities here domestically, how do you do that? And how do you do that in a way that's economically viable? It seems that automation has to be a part of the answer.

And I think our thesis is that if that's the case, we want to have a workforce that's prepared to work with that automation. And we want the automation to be developed in a way that lends itself to being a tool, not just this very simple, straightforward notion of I'll just replace people. That one seems extremely unlikely for many reasons. It needs to be a tool that people can use to improve the quality of their work. So I guess I didn't really answer your question, partly because I don't how to, but partly because I think that where we need to put our energy really is into how do we create tools that make the domestic worker more effective.

CARL MOORE:  Well, in terms of the tools, I think that one of the things that could make the process one in which there's less displacement of workers and more opportunity for workers is getting back to what humans can do. And so most of the tools that we use, whether it be on the job or at home, we use our hands. We hold a hammer. We hold a drill. We use a socket wrench. But when you look at robotics, you don't see hands, right?

So a lot of times, roboticists will talk about hands, but really they mean grippers. On the manufacturing floor, there might be suction cups, but not something that's dexterous. And so we, as part of this NSF Award that you're talking about, we say, well, what if instead the robots had hands, had human hands, anthropomorphic, similar to what we have, and, therefore, that opens up a huge area of possibility where these robots can then do work similar to what humans can do, at least in terms of those same tools.

That really decreases the cost of integrating these robots into different places. And so right now where you have large firms and they can afford to-- well, the robots themselves aren't expensive. It's the integration. It's the changing. It's the programming of those robots. But if the robots have end effectors that are more human-like and more easily programmable-- and we can talk some more about that-- that really opens up them being able to be used by smaller firms where the productivity has been lagging. And so that's one of the reasons we call it HAND is because we're looking to develop hands to be on the ends of these robots.
 
KRISTEN BROADY: So I want to pick at this question just a little bit more. So different industry. I want us to think about cashiers for a second. So between 2019 and 2022, partially because of the pandemic, we lost like 350,000 cashier jobs and like 650,000 retail salespeople jobs. This is not people deciding they didn't want to go to work, or maybe they went and did something else, but what I'm saying is those jobs stopped existing.

And if you went back to the stores after the shutdowns, you were faced, particularly at Walmart, and CVS, and Dollar General, with many more automatic checkout machines. And some people loved them and some people hated them. I am an introvert. I loved them. My mother is 78. She hated him. And now we're seeing that go away, right? Like Target was really invested in this, so was Dollar General. And they've backed off to saying that 10 items or less, or just shrinking the amount of them for claims of theft, customer satisfaction, et cetera.

So just thinking about those changes, I want to apply that to your research and the things that you've been building. Can you tell us about some of the starts and stops maybe that you've had that you develop something that you think is going to work-- like you talked about having to flip this piece up, that you've got this piece of technology-- but thinking about it, and coming to your point, how it really works with humans, what kind of starts and stops have you had that make these things more useful to companies, and to make jobs safer?
 
ED COLGATE: Yeah, well, we can certainly talk about our early work in cobotics. So we actually formed the company Cobotics in I think the late '90s. And so what were we doing? At the time, we were trying to develop a new class of robotic technology that was really intrinsically safe around people. That was the first insight is that, well, if you're going to, at the time, if you found any robot in a manufacturing facility, you found a cage around it, you found laser lights, that if you broke the laser light, everything would shut down, and a lot of training, and keep robots and people apart.

Our insight was, no, there's something to be gained if we can bring them together. And so we worked very hard at addressing that problem. But frankly, our business didn't scale and didn't really succeed. It was only about a decade later with the advent of the first cobotic arms that came out of Europe, that the whole area of robotics really began to take off. So what was the misstep?

And for us, the misstep was that of understanding how to develop something that could scale. So automotive manufacturing, which was our sweet spot, everything has to be kind of perfect in automotive. It's a very involved supply chain. And we were actually working with the OEMs primarily, also, at other points in the supply chain.

And if you bring a piece of production equipment in and it brings down a line that's putting out-- at the time, they were like $40,000, it is a big number-- $40,000 SUVs and you stop that production for one minute, you got a problem, right? [LAUGHS]

And so that meant that the integration of our technology into the assembly cell really had to be perfect. And so that is expensive and slow. And frankly, it often would fail because the production manager just may refuse to take the risk. The brilliant thing that the Danish company that really started building arms-- they're going to be called Universal Robotics-- did is they first, they built a general product, an arm. They still had to put different end effectors on it.
 
But they geared it really only toward small and medium enterprises further back the supply chain and didn't have that risk associated with it. So I think, to me, that's an important lesson is-- also, think about your example of driving the cart around and didn't you wish you had a robotic hand steering it? No, that's actually not the right answer there, right? Because when you can essentially, build the robot into the steering, you do that to get rid of the hands altogether.

So there's always this tension between trying to build something that's general that you could apply across a large set of areas which we tried to do in the late '90s and failed because there was just too much integration, too much specialization to occur. And building a business that's more of a system-integration business, that's really all about the specialization. And so where a technology lands on that spectrum is really critical to its success.

And I think what our theory is today is that robotics-- really due to something we haven't talked yet about, but AI-- is at the precipice of being a much more general-purpose technology than it's been. And so there's a new opportunity to take a look at how these general-purpose, and, therefore, very scalable solutions can be created.
 
KRISTEN BROADY: So I want to think about, again, these starts and stops and how we prepare workers and companies for these technologies. And for me, having been a college professor for 18 years before starting the Fed-- which is the best job ever and I'm never leaving.

[LAUGHTER]

Yeah, never. So I was at Kentucky State University. I was the Vice Provost. And again, I like to have the experience. Whatever the workers, the students, or whoever it is is going to be doing. I need to see it, right? I'm not going to send somebody to do something that I wouldn't do. So I go to an internship at Toyota, Kentucky Motor Manufacturing. And there are robots everywhere that do all kind of stuff there. There's a thing moving, and something else is doing-- I don't what any of this stuff is doing, but there are like no people.

And so trying to figure out what do we say to students? What do we say to professors? How do we get a student ready to go and do this? How do we get workers ready to go and do this? How should companies talk to training organizations, whatever they are, to say this is what a worker needs to know and be able to do, or a student needs to know and be able to do in order to work with these robots?
 
CARL MOORE:  Well, one thing we're doing is we're trying to start off by realizing that we do not have all the answers. So the mechanical engineers don't know everything that needs to be known to make some of the things happen that you're discussing. So as part of this Engineering Research Center, there are engineers, but there are also educators, there are psychologists, there are people, there are economists.

So we bring these people together. And it's something that we-- well, we didn't come up with the term-- but in participatory action research or participatory design. And so our focus is not designing something that we think one day we can toss over the wall and people will use, but we bring users in at this stage at the beginning. So this thing started-- we just got funded in August. And we start from the very beginning with even having students involved as well helping us determine what is needed.
 
What do we need to design? What does industry need? Ed can talk a lot more about the industry ecosystem that we have along with this project. So at the ground floor, we want to have everybody involved, and so we are taking their concerns into consideration at the start. One of the big pieces about this is the democratization of the robotics. Ed mentioned this a little bit in terms of if we look back at what revolutionized the ability for the consumer to use computers. So when Apple came out with the mouse and everything now, everybody can use a computer.

We want to change how people even program and interact with robots. We don't want a company to have to have an engineer-- somebody who's graduated, somebody who has eight years of college-- to be able to program the robotics. We want a person to be able to say, hey-- talking to the robot-- I want you to hold the tool like this. And we want the robot to see it and say, oh, I can do that. We want people to be able to program by talking with the robot.

And AI-- as Ed was mentioning-- really opens up the possibilities for some of this to start to happen. So now we have kids coming out of school without college degrees, they're high school graduates, but they are programming and working with robots.
 
KRISTEN BROADY: So I want to see if I understand this, and I think I do because I've been there and seen it. But I want to say this for the audience. So I'm thinking about ChatGPT. And I can type something into ChatGPT and say create whatever and it'll pull it up. So I imagine you've got sensors on your hands maybe. So you get this thing, and you take it out of the box. You put the sensors on your hands, and maybe when I do this, the robot hand does it too. Is that a thing?
 
CARL MOORE: Exactly.

ED COLGATE: That's a thing.

[LAUGHTER]

That is a thing. Yeah. I mean, I think Carl really described it beautifully, the future that we envision where there's just a very, very seamless interaction between you and the robot. But in that future, there'll be phases. There'll be the teaching of the skills, sort of the parent phase, you know? You got a baby robot. You got to teach the robot how to understand the world and have all the skills. And that phase will probably involve people with really fancy gloves doing tasks that they do really well and gathering lots and lots of data.

I think we all know what does AI need. It needs data, lots and lots and lots of data. We have that. We humans have very conveniently digitized that in images and in text. We have not done that for things like how you use a tool. That's going to have to happen. And so that'll be happening. But then that's at the parent phase. And now you've got your coworker phase where you get that cobot or that robot out of the box.

And at that point, how exactly do you interact with it? Is it just talking to it? One of the things we can do is you can kind of embody a robot. You can't do this with a new worker at your plant. You can't just take over their body and all their senses.

[LAUGHTER]
 
But you can do that for a robot. So is that the right approach? Is it demonstrating and having it watch you? Nobody knows the answer. And that's part of the work that we do.
 
KRISTEN BROADY: So I'm going to ask you to do something that I never do when I give talks, to predict something.

[LAUGHTER]

So I'm going to ask you since I won't do it. But so I'm thinking about how cars are made. And I spent a day at Toyota, so that's basically what I know, and from what I've read. So if you hired a bunch of workers, you would have to train them all and they would all do the things slightly differently, right? So if you had a line of do this, put this together, whatever, you could train all of these hands at once to do that thing, and then the car goes to the next phase. So you could train a bunch of these things at a bunch of different dealerships. I heard someone talking about the 300 car dealer-- right? So how far?
 
ED COLGATE: Isn't it interesting? Every time you use ChatGPT, don't you think it's one thing like it's one mind out there and we're all interacting with it? And so this is to your point. It's like once you've taught one robot how to do something, well, they all know this. It's kind of scary. We are very aware of and thinking about how to avoid the worst dystopian aspects of this.

But, yes, you're right. However, let me also make another point. When we talk about the democratization of robots, what we are really talking about is being able to use robots in contexts where today you wouldn't even think about it. And today they are classic automation. If you have the scale, then you can invest-- just like with the mainframes in the '70s-- you can invest in the production engineering to set up your production lines, the capital expense of the robots, the capital expense of the tooling. You can take the time.
And then once you've done all that-- you've spent several million dollars to set up your line-- well, then you think better just run and crank out the goods, pay itself back. And the economics of that work in some industries. They do not work in all, not in many. So as you go further back the supply chain, you ended up with lower volume, higher mix, doesn't work so well. If you go into other industries outside of automotive, going to food processing, doesn't work so well.

So our vision is that the robot becomes a very different type of tool, not one for automation, but one for augmentation, one that you or I could in the morning say, you know what? Today we're going to be making widget G And I'm going to show you how to do it. And I want you to make 300 of them and then stop. We're good on that. And then we'll talk about what's next. And then what's next comes later in the day or the next day, and then you work with your robot to do that.

Imagine a robot around the home. I don't how many of you have elderly parents who are living at home. My parents are nonagenarians who live in the house I grew up in, and it's really scary. And I really, really wish that they had that sort of support for just quotidian tasks like making meals from something reliable, which the home health aides aren't. So that's the vision that we have. It's a very different type of interaction that will occur.

And honestly, that's the history of augmentation technologies-- when new technologies come along, that people can really use to do their work better. Then innovation occurs, and then things change very dramatically. It's like Jeff Bezos could have just had this vision of, well, we will buy a bunch of indie bookstores and automate the cash registers. That wasn't the vision and nor should it have been.
 
And so again, getting back to, Elon Musk's humanoid robots, maybe that's the right vision, but maybe a technology that lets people in our economy unleash their creativity and have new forms of people and robotic physical automation working together, maybe something richer is going to come out of that.

CARL MOORE:  I guess I would say it's a paradigm shift, that it's very difficult to imagine what could be some of the uses for the future of these devices. We had no idea what was going to be possible with the phones, right? So one of the things we were initially thinking about as we were writing the proposal-- this was coming out of COVID-- and we were thinking, well, what if you had a robot that was soft? Because most of the robots are metal. And you could have it at your aging parent's house where people weren't comfortable yet traveling or being around their older relatives.

But this device could massage your grandma's shoulder. The robot could set the table, or you could step in and control the robot to do something from a distance. So that opens up a whole area of, say, hey, maybe now more Americans can age in place. So it's a lot to consider.
 
KRISTEN BROADY: So I want to come back to the automotive industry because there's a lot that goes into making a car. These are huge places. And so I want to think about investments and capital versus labor a little bit. We've kind of come to this. But if you're going to retrofit whatever is going on now, like if you got some machines, you've got some automation, you've got some programming. And so when some new technology comes along, this hand that's going to replace, or you've got a bunch of arms. The arm can do whatever it can do. But for the end, you've got to have a person.

At this point, the technology only goes so far. And so now you're going to have this hand. And so you're going to retrofit arms with hands that do what people do. I guess my thought is with people, you can train them. You can send them for training. They will quit or you can terminate them if their performance isn't what it is, right? I hate to say it that way. But once you've invested in this technology, it's there. And it's not like a person in my mind. Like you can't just take this one hand off. It's like you would have to invest in a whole bunch more capital.

So how do we think about that balance and keeping up with, hey, I can just go to a school and say you need to start teaching kids X, Y, and Z. At the end of the semester, they know it. And I haven't invested in them yet, right? How do you think that works in terms of building a technology, companies buying that technology, and then it advancing, so the capital versus labor investment? How do you think about that, particularly with automobiles and other huge processes like that?
 
ED COLGATE: Well, a few thoughts. When I was preparing for this, I went back and was reading various articles written by some great economists about how advances in AI and robotics are likely to impact labor. And I think it was Erik Brynjolfsson who pointed out that automation is seductive, that business leaders tend to-- it's a very simple equation, I can replace this person with a machine, worked out the economics of that. They often look pretty good.

But it's not just that. It's policy. And it's really something I think this group can weigh in on. We tax capital at a much lower rate than we tax labor, much lower rate. Why is that? Why are we not investing more? Or why are we not doing the reverse to incentivize augmentation technologies, technologies that let us get more out of labor, as opposed to just replace it?
 
KRISTEN BROADY: It makes me think of like the Cobb-Douglas function-- capital and labor. But I'm an economist, right? So how you decide-- machines don't get sick, there's no workers' comp involved, they're not going to sue you. There are a lot of different things. But also, when I get ready to depart with this person, I can depart with them. I haven't invested in them up front. There's also training. Like so I go back and forth in my mind about which one and what happens when things change.
 
CARL MOORE:  Yeah, that is a very difficult equation. One of the examples I like is that if you say, well, we need this type of work done, let's say, on a construction site. So what do we want to do? Right now what we have is a bunch of workers with shovels. So do we now want a robotic backhoe?

And a lot of people will say, yeah, this is what we want. And it's going to be forever. It's going to be big. It's going to be dangerous. God forbid if something goes wrong and it just decides to dig a hole in the wrong place, right? But what if instead we want backhoes that are human-driven but with automation as part of it?

Now, you've taken that human and maybe now that worker even he can do more. He's more productive, but it's safer. And now that worker, he's higher skilled. So he has a greater chance of job security than the guy did before with the shovel. So I think it's a difficult set of choices. But to me, it argues for automation with the human, augmentation instead of complete replacement--
 
KRISTEN BROADY: Makes a lot of sense.
 
CARL MOORE:  --which is expensive and it will take a very long time.

KRISTEN BROADY: I want to move to competitiveness because we've laid out machines. We've talked about labor. We've talked about how they work together. Now, I guess I'm thinking about labor laws, and who does what, and how it's decided, that you go to school and you can major in whatever you want. There's a GPA system or whatever else, right?

Where in some other countries, people get into industries in a different way. It's decided based on your GPA or what your parents did or however it's decided. So how competitive are we? Or just thinking about what you all are doing and your research, this HAND thing, how does this compare to what's going on in other countries? And then, I guess, I'm thinking about how our labor and education system compares and what impact that has on our competitiveness.
 
ED COLGATE: Yeah, again, we're not economists, but we know a bit about robots. And so we can say a little bit about how we compare in that domain. And I'd say a couple of things. One is the US continues to-- and has for decades-- continues to lead in innovation. Most new technologies, be it robotics or as we're seeing today with AI, they're incubated here. They begin here.

But what we don't seem to do as well, and I do think this probably has something to do with policy, is we don't tend to incubate those technologies quite as well and build up the new industries around them. I talked about we invented cobots here, right here in the Midwest. And yet, we really failed to build up and incubate that industry, that today, every major cobotics company is not in the US. They're mainly in Europe and Asia.
 
This is actually true of pretty much all of the robotics industry as well. So that's an area where, frankly, I think there's a lot of opportunity for us to think about how we take our technical innovations and really translate them into things that can have an impact, not just on wealth, but on things like jobs, and opportunity, economic mobility if you will.

In terms of robotics, the other thing I will say is we're by no means the leaders in using robots. China accounts for 40% of worldwide robotic installations. China's not even the leader in robots per worker. That's South Korea. At four times as many as we have here in the States, China is about, I think, double, and Germany is about double. So it's not just Asia.

So in terms of applying the technology to push the boundaries of productivity, we are not the leaders. And there's China with twice as many robots per person as we have in the US. And China, obviously, is a place where labor is much less expensive. So what's going on there? What is happening with that equation? What are we missing?
Don't know the answer to that.

 
KRISTEN BROADY: I guess in that-- and, again, as everyone has said, not the opinions of the Chicago Fed and not really even opinions, but questions-- I guess, I think about what role do unions play in the US, that so a Union may be pushing for more safety, right? And so maybe there's a technology that would make things more safe so the Union may want that. The Union may also be pushing for higher wages.

So I'm listening to what you're saying about China, South Korea. Are there organizations there? I would doubt if there are such strong labor organizations. I could be wrong. I did spend 5 and 1/2 weeks in China a couple of years ago, so speaking from what I saw. But as I think about our labor system, what role does that play in the making of cars and competitiveness?

Because again, I'm thinking about, all right, so it's going to be safer technology. Workers are going to make more money, which maybe means that they can afford more cars. Just a thought. But again, how does that impact us when we're competing against a country that maybe doesn't have as strong labor laws? So maybe unions aren't pushing for the stuff that makes it safer, but labor is cheaper. So are they using--
 
ED COLGATE: Yeah, I don't know the answer to this. I'm certainly not an expert on unions and their impact on adoption of new technology like robots. But I will say this. One of our team members I haven't mentioned-- so our center is actually five core universities. Carl is at Florida A&M. I'm at Northwestern. We also have Texas A&M, MIT, and Carnegie Mellon as partners.

And one of our colleagues, Ben Armstrong, is faculty at MIT. He runs the Industrial Performance Center there and studies these sorts of issues. So what Ben will tell you-- and I think is he tells me, essentially, a bedrock principle among economists-- is that higher productivity leads to wage gains.
 
KRISTEN BROADY: Mm-hmm. I'm one of those.
 
ED COLGATE: And I remember reading a little, things like two centuries of data have shown that the person who runs the backhoe has higher wages than the person who operates the shovel.
 
KRISTEN BROADY: So that's my research, right? So just quick plug. If you go out there and look, I've got a couple of papers about this. And one of them is that jobs at low risk of automation, that 30 jobs at low risk of automation pay more than twice as much as the 30 jobs at high risk of automation.
 
CARL MOORE: Yeah.

ED COLGATE: Yeah.

KRISTEN BROADY: If you look up my name.

[LAUGHTER]
 
ED COLGATE: B-R-O-A-D-Y.
 
KRISTEN BROADY:
 
Right, thinking about my Google Scholar.

[LAUGHTER]

But yeah, OK, so I'm going to ask a couple of questions. And we've got some questions from the audience. What policy issues could impede on automation and robotics in the future, or what regulatory policies could improve automation and robotics in the future? And I feel like that's timely because that's what we were just talking about.

So I guess when I think of this, I think of various labor laws that say that you must provide safety or workers cannot work in certain conditions. And then there's some automated machine that can do that, or you've got to give workers so many hours, so you can't replace workers with machines. It's like how do you balance those two?
 
CARL MOORE:  That's true. That's true.
 
KRISTEN BROADY: Keep workers safe, but don't displace too many of them.
 
CARL MOORE:  Yeah, well, and there's also you could look at the front end and say, well, what about the incentives? So I think someone mentioned tax incentives. So we may lead in, for example, AI innovation, but we don't want to fall behind if we lose the ability to build and power new AI facilities for some of the learning that's going to be required that Ed talked about.

Financially incentivize some of the SMEs to do some of this adoption which is just greatly lagging. And even some of the places that you find, we use this phrase, robotic hubs where you have really regional clusters of high robot adoption. And you will have them in certain places.

So you'll have them in Boston where they're more so focused on AI and humanoids. You'll have them in China. You'll have them in Germany, as was mentioned. But we need more of them. We need more of places where similar to the ERC where you can have a collaboration between industry and universities. And so that's part of what the ERC does.
 
But you also want industries to be able to connect with the suppliers as well and that talent pipeline. So a lot of these hubs, you'll see they're in locations where they have access to those students that are coming out able to do a lot of this work and develop a lot of these systems. And so I think all of that is a role for government to be-- and it's difficult because government needs to be helpful but not be too in the way so there's that balance. And they don't always do a good job in balancing that.
 
KRISTEN BROADY: OK, so I know we've lost some people. I'm hoping that you'll stay because there is a reception. But I want to just ask a question. How many of you hire people? Before you raise your hand, like you, directly? Are you the decider that you hire students coming out of school or entry-level? I'm thinking about research assistants, whatever the equivalent is for you. Are you a sole proprietor? Raise your hand if you hire people. Raise your hand if you're involved in the hiring process at all. Do you participate in interviews-- raise your hand-- at all?
 
CARL MOORE:  Hmm, most.
 
KRISTEN BROADY: OK, so I figured. And so as I think about this, what can you do to help people get ready? How many of you have ever given a virtual lecture to a college class? Raise your hand if you want to because I teach urban economics on Monday nights partially.

[LAUGHTER]

ED COLGATE: Guest lectures.

KRISTEN BROADY: And plus it got you to raise your hand and pay attention if you weren't. Now, I ask you that, and it relates to this because if you're working for a company that makes something, I'll bet if you reached out to your local university training center and said, hey, my name is so-and-so, I work for this company, we make this or do that, and I would love to come in and talk to a class. You could do it from your office or your house and take 10 minutes or 30 minutes or three classes of mine to talk about what you do. And again, I say that because there are career fairs where somebody from your organization or you, and you set up a table and you've got jobs posted.

But do the students and the professors know what you really want them to know and be able to do? No, because you haven't told them. So that's something that you could do for free. Wherever it is that you want to recruit students from, or that you do recruit students from or people from, and when they get there, they know this set of stuff, and you expect them to know these other things but they don't, just reach out. Just a suggestion. I'm half joking about me, again. I teach urban. But I would just ask you to do that. If you have time, just reach out to the dean, the professor, whoever it is, send a random email, and say I'm available to come in virtually or in person.

Anyway, one way that you can help with this.

So there's a lot of talk about how is labor reacting to automation and augmentation. So at the ports, automation is a major concern for workers. But US ports are largely falling behind in productivity. And so, I guess, for me, this is where I'm thinking about the hand because you still need a worker to do something. But it would make it easier and faster. So it seems like workers would be happy and companies would too?
 
ED COLGATE: Yeah. I must say I don't know much about what happens at the ports. It strikes me that there's been a lot of standardization, shipping containers, and such. And, and that it tends to lend itself toward more special-purpose automation as opposed to the more versatile things that we are working on here. But certainly, it would make sense to me in the vein I was just talking about that you've got to be competitive. You've got to drive productivity if you want to be providing jobs at all, and, certainly, if you want to be providing jobs that command a premium.
 
KRISTEN BROADY: All right. So I'm going to ask this question. I don't know who asked this, but it's definitely not for me.

[LAUGHTER]

And if one of you all can answer this, fine. If not, we'll move on to the next one because I don't know. But how do you look at the value equation between biped robots or humanoids to 6-axis arm-hand robots? I have no idea what that means.
 
ED COLGATE: We love questions like that now. Now, we're getting into our territory.

AUDIENCE: That's [INAUDIBLE].

ED COLGATE: Yeah, absolutely, yes.

CARL MOORE:  I've been waiting for that one.

ED COLGATE: Do you want to take this?

KRISTEN BROADY: Whoever asked that, ask more.
 
ED COLGATE: Whoever, we love you.
 
KRISTEN BROADY: Ask more of those questions.
 
AUDIENCE: So we have the economist interviewing engineers, I'm the engineer. I interview economists. [LAUGHTER]
 
KRISTEN BROADY: I like this.
 
CARL MOORE:  So bipedal versus the arm. I guess the ultimate would be you want to bring them together. There's many cases where just like the human, an arm attached to some kind of backplane that does not move, does not have much reachability. So if you can imagine--

KRISTEN BROADY: So I'm thinking about my television mount. So I've got a TV. It comes out from the wall, and so there's the one that just holds it. Then there's the one that comes out. Mine does this, but it's still affixed to the wall.
 
CARL MOORE:  Right, right, right, right. So the ability-- and it's interesting we've been having discussions about that-- so we envision there being a desire for there to be a torso of some sort. But the research--
 
KRISTEN BROADY: Some sort of stand.
 
CARL MOORE: --and I think Ed earlier mentioned humanoids. That's not the focus. But the ability to move, whether it be on a wheeled base or a legged base-- two legs, six legs-- that's not my area of research. But it stands to reason that you will need this hand to be able to reach things and do things. And to do that, you can't just have the arm on a wall generally.

ED COLGATE: Can I riff on this for minute?

CARL MOORE: Yeah.

KRISTEN BROADY:
 
OK, so just for myself, so what you're saying is my television is now on a stand that I can wheel around, or put the TV on the Roomba, right?
[LAUGHTER]
 
CARL MOORE: Yeah, there you go.

KRISTEN BROADY:
 
And it can follow me or it can-- OK, is that the idea?
 
ED COLGATE: Well, yeah. A hand isn't very much use unless it's on a wrist. And the wrist isn't much use unless it's on an arm, and then trace it all the way back to tailbone and beyond. But so I think for those of you who just follow the popular news and what's going on in robotics, you'll see a lot of videos, especially from this company, Boston Dynamics, that does amazing things, including a lot of humanoid bipedal robots that can do backflips and run and do amazing things.

And then I've already mentioned Elon Musk's Optimus robot a couple of times, and it's very much humanoid. And then there's a slew of these humanoid startups, all of them bipedal. And so the question is why? Why do we put our robots on legs? It's a lot easier to put them on a wheeled base, a lot easier. And the standard argument that's given is really the same argument that we give when people say, well, why make a hand look like a human hand instead of just a simple two-jaw gripper? Like what's the advantage of the human hand? In fact, I've had people say with a gripper, you grab it, now you got it, you know?

But then, of course, all of your dexterity is such that you have it comes from the arm. Whereas, for us, a lot of it comes from what the hand does. So the standard answer to both these types of questions is, well, because we're trying to get these things operated in an environment that was created for humans. So getting up on this podium today, using the steps, or dealing with cluttered floors and so forth, those are all arguments for bipedal locomotion, and then, for us, tools, hand tools, all manner of just manipulations that were built or designed for work with the human hand.

It's not clear that those arguments are really the right ones, though. We may both be wrong. And it may be that the robots that succeed in the world will really never have either bipedal locomotion or anthropomorphic hands. We'll find out. There's some argument that why are we bipedal. And it's probably because we evolved from quadrupeds and then there was this evolutionary pressure to develop dexterity in our hands.
 
And where are you going to get hands? You're only going to get them from your two front feet. And so now you've really got to lift those off the ground and now you're a biped. But if you didn't have that evolutionary pressure, if you had others as many other animals did, well, you'd never give up use of your quadrupedal locomotion system. And so that's the other way that things might end up going is that we'll have dog robots wandering around our factories. It's a fun thing to think about.

CARL MOORE:  It is fun.

KRISTEN BROADY: Well, we are out of time. So I do hope that you will look towards what these two are going to be doing. I'm not exactly sure what it is.
[LAUGHTER]

But robots that hopefully will be accessible to you and your organizations that will not displace workers but cause people to make more money doing exactly what they want to do or whatever the dream is. That's what these two are doing.
 
ED COLGATE: And if I can just say we are very early in this process, and if you have ideas of where or how to direct technology so that it might actually benefit your organization or more broadly, society, get in touch.
 
KRISTEN BROADY: So if you want to talk to these two, none of us, I don't think, are disappearing right now. So if you have questions or thoughts, we will hang out a bit and be at the reception. I want to thank Kristin Dziczek, thank Rick Mattoon. Ed, Carl, thank you for reaching out to me almost two years ago and for coming today.
 
CARL MOORE: You're welcome.

KRISTEN BROADY: Lauren, Peter, thank you for making all of this happen. And thank all of you for sticking around and listening. I hope that you learned something. I know I certainly did. I'm going to turn it over to Rick.

[APPLAUSE]