Where is the automotive industry headed in the age of digital transformation and artificial intelligence?

Of course, we all know what driving will look like tomorrow. After all, Hollywood has shown us as far back as 1997! In the movie "The Fifth Dimension," cars fly through the streets and police cars engage in an hair-raising aerial chase with cab driver Bruce Willis.

But wait - not so fast! Until then, we all still have a long way to go. According to Capgemini, automotive companies lag behind other industries in digital transformation.

Flexera, a provider of IT management solutions, claims in its new ITAM report that only 28 percent of automotive companies have adopted new business models based on digital technologies. At the same time, only 19 percent themselves believe their digital platforms can reach customers who would have otherwise slipped through their fingers with traditional approaches. Less than a third of companies say they have the necessary digital capabilities to manage the digital transformation.

Automotive manufacturers are particularly struggling to use digital technologies in experience design, according to the report.

- Only 40% use social media to improve their knowledge of markets and customers - compared to 47% globally

- 34% use channels and apps to sell products and services - globally, 43% do.

- Only 28% enable their customers to self-serve digitally - globally, it's 46%.

- And only 27% use embedded devices to improve ir knowledge of markets and customers - globally it's 39%. Yet the car itself is the best "embedded device" imaginable.

The automotive industry is also struggling to innovate its business models.

- Only 28% have introduced new business areas based on digital technologies - globally, the figure is 39%.

Yet digital transformation affects the automotive industry at all corners and ends, in purchasing and sales, in customer relations, in safety.

Self-driving cars and trucks are expected to trigger a real revolution on our roads, and AI will tell us when a car is about to break down or an assembly line is about to stop.

But the auto industry may still be the furthest along in introducing digital technologies and AI into manufacturing - no wonder, since the first automobile was built in Germany.

As a young reporter, I worked for the Rhein-Neckar-Zeitung in Wiesloch, and they are proud to have the world's first gas station there.

On August 5, 1888, Berta Benz famously took her husband's freshly constructed three-wheeled Benz Patent Motor Car in Mannheim and set off with her two young sons on the very first long-distance trip to visit her mother in Pforzheim, 120 km away. At some point, she ran out of gasoline, so she stopped in Wiesloch and bought a few liters of Ligroin from the town pharmacist Johann Philipp Bronner, a light gasoline that was mainly used for cleaning and degreasing purposes at the time.

She then chugged happily home, where, according to legend, her husband continued to reproach her for days - after all, she had also been guilty of the first car theft in history - but he later forgave her.

The car of tomorrow will be built in the factory of tomorrow. What it will look like, however, is still unclear. In layman's terms, it looks like a shiny virtual world, with people in lab coats and robots doing all the backbreaking work. Wrong, says Peter Smith of TE Connectivity, a Swiss company with American roots. Tomorrow's factory, he believes, will look just like today's, only it will be populated with hundreds of thousands of sensors, all connected to each other and to central control and AI systems So together they will form an all-encompassing and largely autonomous network - the so-called Internet of Things, or IoT.

But IoT itself is only a first step. The goal is called "Robotic Process Automation," or RPA. IoT sensors provide a huge flood of data, but to forge meaningful RPA applications from it, systems need local intelligence.

Increasingly, sensors, microprocessors and radio transmitters are being combined into a single device. Such highly integrated devices are capable of performing complex computational operations without first sending the data to a central server and the instructions generated from it back the same way.

The intelligence in such networks is therefore shifting more and more to the periphery. Experts therefore speak of "edge computing." Such edge systems work faster because they lack the so-called latency effect that occurs when data is transferred. Even if the delay is only milliseconds, this has a noticeable effect on the performance of machines and control systems.

This is important when using VR and AR in production, for example. Working with so-called data glasses requires latency times below 5 ms, otherwise wearers will feel dizzy. The balance organs in the ears and the information from the eyes do not match exactly. In the worst case, errors can occur, which must be avoided naturally.

In order for industrial production to organize itself autonomously and react flexibly - which is basically what the keyword "Industry 4.0" is all about - huge amounts of data are needed. But unfortunately, most companies, not only but also in the automotive industry, are not yet using this data in a meaningful way.

"On average, between 60 and 73 percent of all data within a company remains unused for analysis," claims Mike Gualtieri, Principal Analyst at Forrester, a market research institute.

Year after year, this costs German businesses billions in lost revenue, lower efficiency and productivity, and avoidable quality problems.

This is often due to the lack of capacity of small and medium-sized companies in particular to process large volumes of data. Yet there have long been systems on the market that can even perform such tasks quite cost-effectively. The best known is probably Apache Hadoop, a distributed Big Data platform developed by Google and handed over to the Apache Software Foundation, a foundation, in 2008. It functions as a voluntary open source organization that, among other things, also maintains the Apache web server, one of the most widely used web servers in the world.

Hadoop is a distributed software architecture that is able to distribute computing tasks in parallel across multiple computers, known as clusters, and therefore process extremely large amounts of data, up to several petabytes, with very high performance. Hadoop-based solutions for enterprises are offered by several vendors, such as Cloudera, MapR, or Hortonworks, which can also help with implementation. Spark is a similar product that is also under the care of the Apache Foundation and has function libraries that provide typical machine learning algorithms.

Such computer systems, which can autonomously, i.e., self-directly identify structures and patterns in large amounts of data, allow companies to identify trends and anomalies - and do so in real time, i.e., during production operations. You can think of this as a crystal ball that helps you see into the future.

It is said that a skilled machinist can tell if something is wrong with a machine just by listening to its noise. The young Israeli company 3d Signals is taking advantage of this ability by evaluating the acoustic signals a machine produces. Their system, called Predisound, consists of a number of ultrasonic sensors installed on the machine being monitored, which can detect slight variations in the sound pattern. Thanks to machine learning, the system gets better and better over time at telling whether a discordant tone signals a failure or not. The maintenance engineer is alerted in time and can repair the damage before it occurs. Fixed maintenance intervals are therefore superfluous.

BMW sees artificial intelligence as an opportunity to improve productivity and efficiency in its assembly plants. The automaker is using the technology for a variety of quality-related applications, such as defect testing and inspection tasks using automatic image recognition and image marking.

As part of a pilot project using smart algorithms, the Bavarian manufacturer aims to improve quality in the paint shop. ensors are to collect data for automatic surface inspection.

BMW wants to gradually replace the camera portals still permanently installed in production with AI-based applications.

The Dürr Group in the Swabian town of Bietigheim-Bissingen shows that it is not only giant corporations like BMW that can successfully rely on digitization and AI. Founded in 1896 as a construction glazier's shop, it is now one of the "hidden champions" and a global market leader in paint finishing technology. The long-established company has opted for a strategy they call digital@DüRR and developed an algorithm called EcoScreen, which accurately evaluates all robot and process data in milliseconds to detect even the smallest deviations in real time. Employees can get to the bottom of the cause of the error before the body even leaves the paint booth

What we learn from all this is that in the automotive industry, as everywhere else, data is the new oil. But in the past, mechanical engineers dominated automotive engineering, and Germany is justifiably proud of its long engineering tradition and of such names as Gottlieb Daimler, Karl Benz, Robert Bosch, Rudolf Diesel, Carl von Linde, Hermann Appel, Karl Friedrich Wilhelm Borgwald, Josef Ganz and Werner von Siemens.

But to handle such huge amounts of data, we need a different kind of engineer, namely software engineers. And unfortunately, they are in short supply in the automotive industry as well. What's more, mechanical engineers and software engineers don't usually get along very well, because they speak completely different languages. This problem, commonly referred to as the "war for talent," will constantly accompany the automotive industry in the coming years, and a lot will depend on how it deals with it and how it solves it.

Which brings us to the next big topic for the next few years: Autonomous vehicles. Self-driving vehicles are robots; able to steer the vehicle themselves with little or no human intervention. Using algorithms, an advanced driver assistance system can use combined sensor data and real-time HD maps to provide a safe and enjoyable experience for both the driver and the occupants of the vehicle. Or at least, that's what we'd all like to hope.

At least since an Uber vehicle ran over a passerby in Arizona in March 2018, the question has been debated among the general public, but especially among insurers: Who should bear responsibility for the actions of a machine?

And to complicate matters further: Who should be liable for a self-learning algorithm? Conceivable candidates would be the programmer or the manufacturer of the program, but also the manufacturer of the car or its owner. Or should we consider this all a case of force majeure, namely digital force?

There's no question about it: autonomous vehicles will increasingly be forced to make decisions about the life and death of people.

In essence, this is not an entirely new question. In a moral-philosophical thought experiment that she called the Trolley Problem, the British philosopher Phillipa Foot tried to prove as early as 1967 that machines are also capable of making ethical decisions, and that they should also be judged on this.

The name is derived from the English term for streetcar - trolley.?Foot wrote: "A streetcar is out of control and threatens to run over five people. By switching a switch, the streetcar can be diverted to another track. Unfortunately, there is another person there. Is it permissible to accept the death of one person (by throwing the switch) in order to save the lives of five people?

I met a charming, if somewhat quirky, philosophy professor named Janina Loh at a panel discussion in Vienna a few years ago, and there the question of built-in ethics in self-driving cars was raised. She laughed out loud and then said, "What ethics would you like - there are so many!"

We spent a wonderful evening at the Heuriger constructing cars that would operate according to the precepts of different schools of philosophy. We drank quite a bit of wine, so I can only remember three.

The Bentham-Bentley would be based on the school of utilitarianism, also called utilitarian philosophy, popular mainly in America, conceived by British social reformer Jeremy Bentham in the 19th century, which holds that anything that "produces the greatest happiness for the greatest number of people " is good. The 'Bentham-Bentley' would make the decision of whom to kill in a flash, assuming he had sufficiently fast access to the Internet, by weighing the potential future benefits that the person in question would represent to society in the future. So: old mother - no. Child in a stroller - yes: He could, after all, be the next Einstein.

?A Kant-Chrysler, on the other hand, would decide according to Immanuel Kant's deontological, or duty ethics, according to which human life is the highest good. Incidentally, this is also the basis of our constitution, which says in Article 1: "Human dignity is inviolable." So the Kant-Chrsyler would probably not decide at all, because every decision would be immoral. He would just go straight ahead and leave the decision to fate.

Whatever ethics we ultimately decide on, we must actively and aggressively communicate these "digital ethics" - in our schools and universities, in training centers, and especially in our companies, where there should be an "ethics officer" in every company in the future, just as the data protection officer is taken for granted today.

There has been a lot of discussion about the pros and cons of autonomous vehicles, so I will be brief here and say what I believe.

I believe that it will take much longer than most think before a car can completely replace its human driver and relegate him to the back seat.

But I firmly believe that we will soon see self-driving trucks all over the world, because this technology already exists and has already proven itself - albeit only on the long haul. But from a warehouse in Hamburg to a distribution station on the outskirts of Munich, such vehicles would be of great benefit, especially since everyone complains that there are no more truck drivers.

The same goes for self-driving buses in pedestrian zones, which has been working for a long time. And autonomous cargo ships are no longer dreams of the future either. So maybe we should rather focus on such applications for the near future and continue to drive our cars ourselves. My opinion...

As far as cars are concerned, I think new business models like shared mobility are considerably more promising. In so-called ride-sharing, after all, a means of transportation such as a car or bicycle is shared among users as needed, which contributes to less pollution and road congestion.

Cloud-based applications based on shared mobility solutions can be driven by live navigation with up-to-date traffic information or alternative routes, and show routes with the least traffic along with estimated time of arrival. There will also soon be algorithms for ride service providers like Uber or Lyft that will alert the driver closest to the customer, which can reduce idling and waiting time.

Modern cars are nothing less than supercomputers - or like an iPhone on wheels. They generate reams of data from which real-time alerts and information can be derived about things like tire pressure, GPS, temperature and more. This data can be stored in the cloud while the vehicle is running and later used for telematics, infotainment systems or vehicle status displays, advanced driver assistance systems (ADAS) and mobility services.

Even software updates can already be delivered wirelessly via the Internet directly to the car - no more workshop visits necessary!

This goes so far that the head of TüV-Süd, for example, with whom I spoke about this some time ago, can imagine a world in which car manufacturers can collect so much data about their vehicles that they will be able to certify their roadworthiness. And where does that leave TüV? A business model that is more than 120 years old could disappear overnight!

Most importantly, such a connected vehicle could provide deeper customer insights. The resulting data could help derive trends for manufacturing, supply chain, marketing, predictive QA and customer behavior.

But the car of tomorrow will also serve as a means of payment, quasia as a driving wallet: With connected vehicles, payment methods can be integrated directly into the car. Drivers pay for mobility services directly where they need them, for example, to activate functionalities according to the pay-per-use principle or when refueling, parking or using drive-in services.

BMW, Ford and three others have already introduced blockchain-based cashless payment systems. The idea is to enable self-paying cars that allow drivers to automatically pay for parking, highway tolls, gasoline and other purchases without having to use cards or cash.

But it doesn't stop there! Shopping by car is the new buzzword - and all without having to get out of the car!

The average American spends about 46 minutes a day in their vehicle. According to Mark Lloyd, Consumer Online Officer for General Motors OnStar, work and home are usually the most important destinations we want to reach in those 46 minutes. And the second most important destination? Dealerships.

During those 46 daily minutes, Lloyd said, we mostly visit the same places. Only 5% of the time do we use navigation, which means we want to go to a new place.

Therein lies a huge opportunity for retailers. They can engage drivers directly before they reach the store, or before they even know they're going to drive to the store.?And they also offer the opportunity to provide new shopping experiences to drivers as a service.

What might that look like? Let's look at a couple of typical examples.

- Refueling. Your car detects that you're out of gas, but instead of simply lighting up an indicator on your dashboard, your car directs you directly to the nearest gas station. Because you've stored your payment information, the pump is automatically activated as soon as you pull up, and your account is automatically charged for the amount of gas. In the U.S., Shell has already announced plans to enable such "payment via dashboard."

- Getting coffee. Every morning on your way to work, you stop at your favorite coffee shop and grab a coffee for the ride.??Your car detects that you're on your way to work again based on contextual information like your route and the time of day, and makes recommendations about what you might order.?Normally you go for a hot coffee, but today it's particularly warm outside, so it suggests an iced coffee instead. You choose, and the order is placed right away.???

- Diaper shopping. You're a busy parent and you keep your shopping list on your smartphone. The syncs the shopping list with your car, and that reminds you on the way home from work that you need diapers. You place the order right from your car, and a store associate receives a notification to fill your order. When you arrive, the employee is notified and brings the order to the curb for pickup.

Of course, the connected car is also causing new problems. Tesla made headlines the other day because Chinese authorities banned vehicles from the luxury brand from near military installations. The dozens of cameras could be used for espionage purposes.

But it's not just the Chinese who are worried: the security officer of the Berlin police announced in June that Tesla cars would no longer be allowed to drive or park on the authority's property in the future - for example, if you want to make a quick trip to the authorities.

Clearly, the automotive industry is going to have to pay a lot more attention to cybersecurity than it has in the past.

Car companies are either consolidating into huge conglomerates like Stellantis or shrinking and focusing like Mercedes. Much of this is due to the huge shift to electric vehicles and cars essentially becoming rolling computers. The new CEO of Mercedes-Benz, Ola K?llenius, often refers to cars as "digital products" - and to Mercedes itself as a technology company. (Actually, he says it's a luxury and technology company).

A fundamental barrier to electric vehicle adoption is consumer confidence in the availability of public charging stations - a key enabler of electromobility.

Some scholars argue that centralized public investment in e-vehicle infrastructure would be pointless in the long run because consumers will increasingly have access to charging at home. Others argue that public spending on e-vehicle infrastructure is "wasteful" because it would go toward infrastructure that will be built by the private sector anyway.

However, there is no evidence that privately operated charging stations outperform publicly operated or managed ones, such as government or municipal buildings, public libraries, rest areas, and public parks, in terms of audience favorability.

But the prize question remains: will people want to pay for all that Digital Transformation offers them? Doubts are warranted here. According to the Deloitte Global Automotive Study, between two-thirds and three-quarters of all consumers surveyed are unwilling to pay more than $500, or about €400, for new technologies in their cars. Funnily enough, this unwillingness is particularly strong in Germany - supposedly the number one car country in the world. Koreans, Chinese and Indians are much more willing to dig much deeper into their pockets to acquire the car of tomorrow.

We started today with a wake-up call. Digitization, connectivity, mobility and artificial intelligence are changing the world in more ways and much faster than anyone could have imagined just a short time ago. My friend, futurist Gerd Leonhard, claims in his book "Technology vs Humanity" that humanity will change more in the next 20 years than in the previous 300.

So the question you should be asking yourself is not, "Will change happen?" Rather, "Are you ready?"

But whether we like it or not, change will come. All we could theoretically do is hold it up - at least for a little while.

But that's not a good idea - not for you, not for your company, not for society - and certainly not for the automotive industry - on which we all live.