Episode 1: AI in Automotive

Anni Tabagua: Welcome to Biz Tech Forward, your go-to podcast for cutting-edge insights at the intersection of business and technology. Join us as we explore the trends, innovations, and strategies shaping the future of the digital world. Let's move forward together. Hello and welcome to Tech Forward, the podcast where we delve into the world of technology and business with some of the brightest minds at DataArt.

I'm Anni Tabagua from the Media Relations team, and I get to work with these brightest minds every day. So, please think of me as your friendly tour guide as we discuss the past, present, and future of tech. Today, we're kicking off our very first episode with a fascinating topic: AI in automotive. So, we're literally going to talk about how AI is driving the future. Our guest today is Dmitry Bagrov, the managing director of DataArt UK.

Dmitry Bagrov: Hi, Anni. It's a pleasure to be here. If you could introduce me to those bright minds, that would be very useful.

Anni Tabagua: Dmitry has loads of experience running the software company and dealing with clients, and he often comments on tech trends for various media. So, Dmitry, again, welcome to the show. Thank you for being here. We will try to break our discussion down between the past, present, and future. So, Dmitry, let's start with the first. Obviously, how did we get here? Would you give us a little background on how we arrived at the current trends in automotive? What key developments would you highlight?

Dmitry Bagrov: Certainly. So first, the Earth cooled down, then the dinosaurs appeared, and then the meteor struck. Then, the dinosaurs died, and they turned into fossil fuels. And that was the beginning of the automotive industry. However, AI in automotive is not a very recent invention. Artificial intelligence has been around for quite some time now, at least in theoretical form.

The first types of AI appearing in cars date back to the 1950s. Realistically, the first artificial intelligence introduced into cars was cruise control, and it was obviously getting better and better. As the computational power of the chips that could be used in cars increased, the cars got ABS, anti-lock braking systems, and various driver assistance systems like lane assists, which also date back more than 30 years to the early 90s.

However, the real sort of AI, the real game changer, came in the early 2000s when computer vision and data processing allowed more sophisticated equipment to be installed in cars. So things like adaptive cruise control and lane-keeping assistance, where the computer analyzes visual inputs and is capable of making decisions based on that input, were real game changers.

DARPA, an American agency, as far as I remember, had a grand challenge in 2004 where they wanted to showcase the potential for autonomous vehicles. That event gave a big push to all things AI in the automotive.

Anni Tabagua: Oh, in 2004, like 20 years ago.

Dmitry Bagrov: 20 years ago. Yes.

Anni Tabagua: And that is the DARPA Grand Challenge.

Dmitry Bagrov: DARPA Grand Challenge. Google it up. I couldn't be bothered.

Anni Tabagua: Speaking of challenges but other kinds of challenges, would you also comment on some of the actual breakthroughs and challenges that have influenced the development of AI in automotive?

Dmitry Bagrov: I think the biggest challenge is actually something I already mentioned: it's the analysis of data, visual or environmental data. If we describe it in a more general sense, vehicles are now equipped not just with cameras.

They also have lidars and radars. All that data needs to be analyzed, and decisions need to be made based on that. That includes very fast data processing. You should be able to operate or store data on the vehicle because you obviously cannot rely on data transfer for split-second decisions. Machine learning, particularly deep learning, plays a very important role.

All of that took us to the point when regulatory organs and bodies started messing around with autonomous driving, which is actually a good sign because, from my point of view, whenever regulations appear in something, it means it's mainstream enough to be dangerous. So it's a good sign. While autonomous driving was confined to labs and test tracks, there was no need to regulate it.

When it spilled out on the streets, obviously, there was a need to regulate it and create some certifications and other nice bureaucratic things. Another breakthrough was actually spurred by not cars but more small devices at home, which is basically IoT, the Internet of Things. There is a need to get data from other cars or other players, let's say, on the roads.

It was always there, but once it was clear that the methods used in smart devices at home for communicating with each other could actually be used for cars to get information from their environment, that was another major breakthrough. Of course, it meant additional data. And, of course, it meant the need to drastically change the way cars are viewed by their manufacturers.

Previously, a car was mostly an engine with a chassis and wheels and autonomous driving elements with two legs and sometimes two arms somewhere in it. Now, it's actually a software-defined vehicle, defined by the software that runs on it.

Speaking about AI, one of the things that may be less noticed is the growth in computational capability. So the chips that are currently used in cars—I think I had the numbers; I read them somewhere recently—I think where previously it was 30 to 60 different chips, different little computational units in the car doing all of this stuff around it.

Now we're down to 4 or 5. Yeah, that means the capacity grew significantly. This opens up a lot of interesting possibilities and a lot of very scary possibilities as well.

Anni Tabagua: You know what, I wonder? You did mention sales, and it is still data that we are both kind of related to and kind of sort of. I just wonder, say somebody helps you on the street in London somewhere, and they just randomly say, hey, Dmitry. But, quick question. How are companies like DataArt contributing to such advancements? What would you say? I'm just curious.

Dmitry Bagrov: Well, the first question would be, who are you? And how do you know me? Because I'm not a celebrity and people don't stop me on the street. But what we do, I think, is one of the most interesting things we can do, and actually do with some of our clients, is bring the understanding of how the wider IT infrastructure should be created for certain things to work properly.

If you look at it, something very tectonic is currently happening in the automotive industry. I did start with dinosaurs and then the meteor. So AI is a meteor crashing the dinosaurs at the moment because the automotive industry is very conservative. A new model's development cycle is calculated in years if not tens of years.

If you see a new model going out, it means they started developing it 5 or 6 years ago.

5 or 6 years ago, in IT terms, is ancient history. So we are currently on iPhone 15, just about to be, and iPhone 16 is about to be launched. Okay, so let's say iPhone 15, five generations back. That's the iPhone X. Very different things. Aesthetically, computational power, cameras, and a lot of things are very different.

So five years is a very long time in technology. What is happening in the automotive industry now is that they need not just introduce more technology into their cars. They actually need to change their entire process and how they do things. They can no longer, for example, develop the engine, the chassis, the interior, and the exterior for four years.

And then the last year, just tell the software developers, there you go. That's the car you're going to be powering. Just write the software. It doesn't work that way anymore. So, it has to be involved in the sense of the word. Technology has to be involved from the very beginning. The cycle has to be much shorter.

The Chinese manufacturers understand that really well. They churn out new models not every five years. They do it much quicker. So the Western manufacturers, the European, German, and American ones, are actually facing, I would say, an existential crisis at the moment because, on one side, they face competition from the Chinese, who can do the same that they can do, but ten times faster.

On the other hand, they can no longer continue doing what they have been doing because everyone is expecting different. So, I think DataArt can be very useful because we understand the overall technology life cycle. We understand how technology should be approached as part of a larger thing so that this works.

Anni Tabagua: Yeah. And we are fast.

Dmitry Bagrov: Speed is not the most important quality in development. Any builder will tell you that.

Anni Tabagua: Because the most important quality is?

Dmitry Bagrov: The most important quality is quality. It has to work, as it is only a binary system—ones and zeros. Either it works, or it doesn't.

What a lot of people don't get is that something that worked before in one context, taken out of it, and moved to another environment wouldn't work. So it's very subjective. And that means, I mean, that's really good news for us because it means even if we did something, we might do it again and again in a different environment.

Anni Tabagua: Well, not to brag too much, but, at DataArt, it works.

Dmitry Bagrov: Most of the time.

Anni Tabagua: This was a good little bridge to, I think, my favorite part, which is about the future. So, I would like us, or I would like you to predict the future. Please. Where are we going for?

Dmitry Bagrov: Which version of the future do you want? The utopian or dystopian?

Anni Tabagua: The optimistic?

Dmitry Bagrov: Optimistic? Okay.

Anni Tabagua: Realistic.

Dmitry Bagrov: Okay, let's stick with moderately optimistic, then.

Anni Tabagua: Okay.

Dmitry Bagrov: As I said, we're at a point in the automotive industry's transformation, and if we give it another five or ten years, it will be very different. First of all, we will see a lot more autonomous vehicles, particularly in cities or situations where the route is predictable.

So buses, taxis, and, to a certain extent, trains—don't forget about trains because, as the past has proved, you can actually have autonomous trains. So we'll definitely see a lot more of that. We will see a lot more alternative fuel type vehicles or electric vehicles, obviously.

Maybe hydrogen cells, something else, because fossil fuels are a finite resource. And it would be good to have something else available. And again, I think the ideal testing ground for things like that is public transport. Well, if you think about how many diesel engine buses are in London and what kind of pollution they make, even with all the regulations, and if you replace all of them with electric buses, it would be a big win.

And then buses are very predictable. They have a route and a depot, and they start at the same points and finish at the same point. They pretty much do exactly the same thing all the time. They need to react to traffic. That's true. But, as Tesla and other manufacturers, such as BMW and Audi, have already proved, the reaction is not really impossible.

So that's, you know, that's one thing. What else?

Anni Tabagua: I think anything that comes to mind that you think might cause, like, a real revolution or something.

Dmitry Bagrov: Some revolution.

Anni Tabagua: The wow effect, you know.

Dmitry Bagrov: Aliens are coming to Earth.

Anni Tabagua: A revolution related to AI and cars.

Dmitry Bagrov: Aliens in little cars are coming to Earth.

Anni Tabagua: Okay, good. Yes.

Dmitry Bagrov: But no, I think processing power is important. If quantum computing finally goes mainstream, apart from all the horrific problems it will create for cryptography, it would allow for much faster computation on board. So that would be a major driver for autonomous vehicles, for vehicles controlled by flying cars.

Back to the Future promised us flying cars—what, 30 years ago? Nowhere near. But I think small aviation is neglected. And it's quite unfair because if you think of it even in the UK, if you go back to the 1950s and 1960s, small aviation back then was a lot more developed.

So small air taxis were around. People could fly from Birmingham to Oxford. I don't know; how long is that? A couple of hours? Driving would take you at least four. And so I actually think that the most interesting application for autonomous driving and AI on boards is not the vehicles that would be driving on the ground.

It's the vehicles that would be flying. Well, I read an article some time ago that tried to analyze why small aviation didn't take off, so to speak. The answer was that, as traffic grew, humans just were not reliable enough to think very quickly in three dimensions, you know, in two dimensions.

Yes, and confined within the road. Yes, and we still have a lot of accidents. Unfortunately, we still have a lot of people dying on the roads. But add another dimension to it and remove the roads, the confines of the actual tarmac, and you have a disaster waiting to happen. So the AI on boards is actually, I think, not for the ground vehicles.

It's there for the flying cars and whatever else will be flying chairs with cubicles and flying meeting rooms.

Anni Tabagua: We shall see.

Dmitry Bagrov: You know, hopefully. But that's not five, ten years. I think that's at least 10-15.

Anni Tabagua: We shall see. I did say this part was going to be my favorite, but I take it back because this next part is actually my favorite.

What I mean now is that it's like this quick bonus segment. Let's say it's a bonus, but to spice things up a little bit, I want you to give us and our listeners three unpopular opinions that you have about AI and automotive. Go.

Dmitry Bagrov: You always ask me to do the nicest things. Okay. Cars will make us lazy. Autonomous cars will make us lazy, and we will all turn into people from Wally, you know, the morbidly obese fat people who don't do anything themselves and drive around in little tiny chairs. And, I think, as apocalyptic as that looks, I am pretty sure it will happen, but probably not on a massive scale.

As always, with any invention, there will be people who just take it to the absolute maximum, and you can't really do anything with it. My personal favorite is that the AI will actually take a proper backseat driver position in your car, expanding from taking the next left to forgetting your breakfast.

It might actually happen because if you look at the way the assistants, various AI assistants, are evolving, the way, for example, Chatty Betty Ford can talk, simulate emotions, make pauses, and stutter. It actually can fool pretty conveniently, pretty convincingly, anyone that they're talking to a person. So, I think we're not yet at the stage when AI will pass the Turing test, but we're very close, and I'm pretty sure that AI will just stop nagging you at some point.

You're going too fast. Slow down. You already had three fines this month. Oh, you're going over your budget. It saturates at two, etc., so there should be a kill switch. I should be able to switch it off. But I honestly don't think that realistically is going to happen. Maybe as a joke, like, you know, Tesla can play something stupid using the car horn, but not on a massive scale.

What I think is definitely going to happen is that AI in cars will be just another channel for sales. So, your car would basically be just another source of data about you. And it will be able to transfer that data to whoever pays for it. And it will be selling your stuff, so it will be selling you.

I don't know if it's the latest vacuum cleaners from Dyson, ergonomic seat cushions, or yet another charger for your iPhone. Nobody needs billboards when you actually have your dashboard as an advertisement space, and unfortunately, I think that is definitely going to happen.

Anni Tabagua: Oh no, we cannot afford to buy more stuff, Dmitry.

Dmitry Bagrov: Oh, if you cannot afford to buy more stuff, don't buy more stuff. But you will still be bombarded by the advertisements.

Anni Tabagua: Well, Dmitry, thank you so much for talking to us today. I enjoyed our conversation; it was really interesting.

Dmitry Bagrov: Thank you. Always a pleasure.

Anni Tabagua: Thank you to our listeners for tuning in to Tech Forward for our very first episode. It was so exciting. Please stay tuned for more because more is coming, and we also want to hear from you. What are your thoughts on AI and automotive? Do you have a popular or unpopular opinion of your own? Please send us an email at techforward@DataArt.com.

And in the meantime, drive safe. Thanks for listening to Biz Tech Forward. Be sure to subscribe and rate the podcast to stay updated on the latest in business and technology. Join us next time for more insights and forward-thinking discussions presented by DataArt.