The rise of electric powertrains, sharing services, and the anticipated arrival of self-driving cars are shaking up the auto industry. Ken Kelzer, General Motors’ vice president of global vehicle components and subsystems, is at the center of that. He oversees the development of every part of every new GM car. Digital Trends caught up with Kelzer in New York City, where he judged a competition for MIT Solve, an initiative that seeks to apply tech to the world’s problems. He explained why a GM executive is interested in MIT Solve, and how he’s trying to prepare a century-old automaker for the future.
DT: Could you tell us a little bit about the MIT Solve competition and why you’re involved with it as a representative of GM?
Ken Kelzer: We are trying to attract, retain, and develop new employees that are coming into General Motors. It turns out that the demographic in our company is such that we’re turning over. In less than five years, 40 percent of the folks who have taken a STEM-related job at General Motors — scientists or engineers — have been new to the company.
At the same time, the industry is changing dramatically. The conventional way we’ve done vehicles for 50 years, and the type of portfolio, “hey here’s your new sedan, here’s your new truck” is going to “here’s your new electrified vehicle, here’s your new autonomous vehicle, here’s your new connected vehicle.” The intersection of that is taking place right now.
For example, Livox was one of the winners. This is a guy from Brazil whose daughter ended up with cerebral palsy seven or eight years ago because of a medical issue that was induced at the hospital; she wasn’t born that way. She has the inability to communicate verbally. This individual is developing an artificial intelligence platform for people to understand [each other] through nonverbal communication.
You say, “why would we be interested?” Well, when you get into an autonomous shared vehicle and there’s no one in it, there might be an element of [wondering] how would we understand and be able to send and receive information back and forth? And if we’re in the personal mobility space, again, how do you communicate to multilingual people with a rationalization of what their structure is as they come into it, hence the connection.
Ultimately, we’re going to get exposed to people that think in an innovative way. We might be able to tap into the kind of knowledge we never thought we’d be able to tap into.
It sounds like you’re going far beyond normal automotive areas.
I’m not trying to move away from the conventional, because I’ve got a large population of engineers who are changing, becoming software where we weren’t, and looking at electrification. As you know, that’s not your typical propulsion systems of 10 years ago.
Will we be going to campuses, and will we be interviewing? Sure, but the world is changing so quickly that we’ve got to figure out other ways of getting innovation in, and that may be unconventional.
GM also recently hosted a Bluetooth hackathon involving MIT students. Can you explain how that worked?
As the engineering lead for GM, I have to go out to all of these campuses. Universities are looking at how they can change themselves, and how they can be relevant to corporations. But what was interesting, and I’ll kind of compare and contrast the old model, the old model is, if you want to do research, you come with money, you say “let’s go do a project,” and the project lasts two to three years. Well guess what, the industry pivots this way and that way.
Now there are campuses like Carnegie Mellon. They have built a dormitory of sorts where they bring industry onto their campus. And I was like, “what happens if I bring someone from my team out to campus?” I have an individual that graduated from MIT that I’ve embedded on [MIT] campus that is networked in everything that’s going on.
We said, “Bluetooth connectivity: kind of an issue for a lot of people.” Not just us as an OEM but, if you read J.D. Power, sort of universally people have a problem connecting their phones.
So, we do this hackathon, we have these teams come forward, we judge them, and we pick this team of four individuals. It turns out that they were competing against PhD and Masters students. The four students that won were all freshmen, and they had never really coded in this space. We brought them to our tech center in Warren for a project to actually work on it and then we hired them as interns. It turns out they had never connected a phone to a car. And they had no one in their family that ever did.
We want to go to campus with problems. And we’re doing more of that this upcoming season.
What are some of those problems? What did that group of interns come up with, exactly?
They’ve actually backed up the interface that we would use on it. We’ve had them work within our team to figure out “why do you go to this screen, I wouldn’t look for that, I would look for this.” Well, that changes our human-machine interface, and now we’re trying out their changes to see how adaptable people are to them.
You mentioned having to keep pace with the development of technology on this research, or problem-solving, level. What happens when you actually start making cars and have to deal with suppliers? The lead time for developing a new car is longer than for components like sensors and electronics, how do you deal with that?
You said it. The timeline to develop a vehicle can be four to five years, and when you insert technology, because phones are changing yearly, it can become irrelevant after they change the technology.
Internally, we are in-sourcing some of that software and hardware development. We think that’s going to give us a competitive edge in the future. Because to do these vehicles—whether they’re electrified, autonomous, or connected— the difficulty is integrating that technology back into the vehicle, and having the ability to have the vehicle stay the same over the four- to five-year [timeline], but only having a one- to two-year time frame to change the technology.
We’re bringing that in-house. We have a software team that’s north of Toronto, in Markham, to do that specifically so that we’re not paying suppliers. But that drives our hardware technology. We’re developing hardware such in a way that we can update it and apply the software to it, and then ultimately do it over the air and have that connectivity.
Speaking of hardware, General Motors relied heavily on LG for help developing many high-tech components for the Chevrolet Bolt EV. Are you saying GM won’t rely as much on outside suppliers as it did with that car?
It’s a combination. We have sort of an internal calculation, I would call it, to figure out whether we want to more vertically integrate and bring some of that technology in. But then there are other areas, especially when the things are so costly. The interesting thing is that the entire industry is faced with the same thing, the cost of some of these components.
In some cases, you just can’t cut the cord. We’re still going to make trucks.
What we try to do is, those [items] that we think are going to be competitive and IP-related, we want to own that. Because we think that’s where the industry value is at. Then there are other areas where we want the Texas Instruments and the LGs to turn it into a commodity so that everyone uses it.
Right now, automakers go through a lot of effort to differentiate cars by segment and by brand. Do you see any of that changing as this new methodology, and new technologies come into play? Will all cars become more similar?
Yeah, transportation as service.
If transportation becomes a service, and you don’t own the car, it’s more commoditized.
I’ll maybe say the question my way, the way I get asked, and that is, if I pick up a Lyft here in 10 minutes and it shows up to be an Infiniti QX60, do I say “it’s not a [Chevrolet] Suburban so I don’t want to get into it,” kind of thing, or am I just looking to get transportation?
I think both hold true to that. I think there is still for many years going to be that identification on brand awareness and what a vehicle is. People feel proud, because vehicles are very passionate for a lot of people, and they identify with it. At the same time, when you get shared autonomous vehicles, I think you get less and less [passion] down there.
I think the things that are going to be the differentiators are going to be the technologies that you put into it. Like, when you get in, how fluid is it to kind of walk in listening to a phone call, and then it’s already in there. Or we recognize you through your app, and then you have some preferences, and we allow you to have those preferences. Or, through MIT Solve, we recognize through some facial recognition the things that are going on. You might be confused, so we help you out.
We make it more frictionless, and it might be a frictionless drive versus [something] more difficult to use, one brand versus another. So, the integration of the vehicle, and that experience, we do think is going to be a differentiator. We need to stay on both sides of that.
How do you manage all of these changes? How will GM transition its workforce and manufacturing operations from what they are today to the future you’ve discussed?
Welcome to my daily life. I think like in anything where you’re trying to make a transition, in some cases you just can’t cut the cord. We’re still going to make trucks. We haven’t even done our whole plethora of our new trucks. You sort of have to continue that aspect of it.
But in the new areas, I think you’ve got to showcase the difference and those locations. So, as an example, I told you we’re transforming the tech center. Our autonomous and advanced work for what you just described as those new technologies. They now are in our very first building from 1955 at the tech center. It probably was our most run-down building. We renovated that and transformed it, and now the new technology comes in.
We’re moving those people in, we identify them as a core group, in a building that says “this is the group that has to do it.” You give the folks the ability to fail. And you get a few less boundaries on how you want to operate, whether that be project management, whether that be failing. So, you allow that to happen, and at the same time I would say the core group of the engineering team weaves into those conversations and makes sure we don’t go too crazy and fall off.
- GM autonomous vehicle chief engineer wants to make disabled access a priority
- Everything you need to know about autonomous vehicles
- Volvo has created an autonomous truck that looks like a sports car
- Cadillac drivers can now find and pay for parking from their dashboards
- MIT and IBM’s new A.I. image-editing tool lets you paint with neurons