Michael Würtenberger, head of “Project AI” at the BMW Group, talks to newspaper DIE WELT about artificial intelligence and how this key technology is already being used across the company. 

WELT
Mr Würtenberger, the “Covid-19 pandemic” is forcing many carmakers to make drastic cutbacks. Is your “Project AI” at BMW affected?

MICHAEL WÜRTENBERGER
No, we recognised early on that AI would be a key technology in the digital transformation. We don’t want to hold back further development of such an important topic.

WELT
The use of AI plays a central role in the digital transformation. It requires an enormous amount of capital, but money alone is not enough. Training and, in the early stages, even availability of qualified employees is a very important part of this multifaceted process. There’s a big shortage of IT specialists...

MICHAEL WÜRTENBERGER
For BMW as a technology company, the skills of our employees are crucial. We need to recruit highly-qualified people to join BMW and then retain them, by offering attractive conditions – from a collaborative environment to a good corporate culture. But enabling our own employees is also important. The so-called digital world is slightly different to that of traditional mechanical engineering. We are trying to reconcile these two worlds. On the one hand, the car is a classic core product, but at the same time, it has also become heavily digitalised and highly complex. The same applies to producing it. For now, neither side can function all on its own. That’s why connectivity within our company is just as important as with other firms.

WELT
Vehicle engineers specialised in combustion engines need to speak the language of network administrators and programme developers – and vice versa.

MICHAEL WÜRTENBERGER
Yes, they need to speak to one another at the same eye level. That is important. In the digital transformation, IT specialists and mechanical engineers have to be on equal terms. We need to learn how to do that, how to cultivate it and be able to handle it. And that is not just for employees – it is also very important for managers.

WELT
You are competing for highly sought-after IT experts against the big technology and internet companies. What can your “Project AI” offer people that Google or Amazon can’t?

MICHAEL WÜRTENBERGER
Real jobs – where success results in physical, intensely emotional products that can actually be built on a large scale! We aren’t just dealing with concepts, but with the concrete challenges of how to use AI technologies like natural speech recognition, computer vision, machine learning and many others to bring a better product to customers. It is really exciting when you have three years to solve a problem and offer customers a solution.

WELT
AI is considered a key technology for Germany’s key industry, vehicle production. How is AI anchored in the BMW Group organisation?

MICHAEL WÜRTENBERGER
AI plays a role virtually everywhere from production to inside the actual vehicle – that's why we have developed a hub-and-spoke organisation. “Project AI” forms the hub and individual business departments are the spokes. This creates transparency and enables know-how to be transferred quickly and efficiently. AI can do a lot of things, but using it in a real process requires considerable know-how and experience from the employees on the ground. That’s why we decided on this decentralised form of connectivity in most cases.

[…]

WELT
You are also working to simplify the human- machine-interaction.

MICHAEL WÜRTENBERGER
One way to achieve simplification is by pre-conditioning vehicles on an individual basis. The car needs to be able to recognise who is using it and which music, interior lighting and seating position that person prefers. We want to make complex operating procedures simple: If the system knows you are going shopping in town, it can guide you to a parking space – taking into account whether it might be raining, and whether you are someone who is prepared to pay for garage parking or would rather walk the last 500 metres and stop off somewhere for a coffee first. If AI makes this easy, you’d probably say: Hey, that was a cool experience!

WELT
AI plays an especially important role in production.

MICHAEL WÜRTENBERGER
One area where we use AI is quality assurance. For inspecting paintwork or weld seams, AI is far superior to classic image processing. There are a growing number of areas where more complex tasks can no longer be managed by conventional rule-based automation. AI enables more robust solutions in these cases – and robustness is ultimately what it is all about.

WELT
Can you put a number on the productivity gains from AI?

MICHAEL WÜRTENBERGER
That is impossible to say in general terms. We have demonstrated positive feasibility for more than 400 different AI and data-driven applications. Productivity gains might only be a few percentage points in some cases, but much more in others.

WELT
Many people are worried the use of AI will put their jobs at risk.

MICHAEL WÜRTENBERGER
That’s not the case at our company. We have a lot of support on the employee side, because we are automating things that are very repetitive or strenuous. It's not about replacing people with machines, but about reducing the strain on employees. To this day, BMW has not let a single employee go as a result of using AI.

WELT
Developing AI applications is complex and expensive. Are carmakers working together in any way to save time and money?

MICHAEL WÜRTENBERGER
When we develop products, we attach a great deal of importance to acquiring core expertise so we can create intelligent products ourselves that give us a unique technological edge. But collaborations clearly make sense in many areas. For example, we are working with companies like Daimler, through the VDA Flagship Initiative for Autonomous and Connected Driving – in particular, on driving standardisation of AI applications. Beyond that, many basic AI components are used through platforms by tech companies like Microsoft, Google and Amazon. We are ultimately taking an open-source approach to AI algorithms, which we published on the Github platform in autumn 2019. We are choosing to work very closely with IT and the AI community to further develop the algorithms our applications are based on. That is very helpful. Our explicit aim is to make using AI as simple as possible, so all employees are able to select the ideal AI tool for their purpose from what is basically a digital toolbox.

WELT
Where do German carmakers stand on AI usage compared to their international competitors?

MICHAEL WÜRTENBERGER
I think we are doing very well. I tend to get a bit restless, so I would always like to be doing more, but I’m pleased with the speed of implementation for the most part. You could always do things faster, but you have to remember: Every technology we push must be accepted by our customers and employees – and ultimately deliver a return on investment for BMW.

WELT
The automotive industry is facing the biggest upheaval in its history – and the digital transformation will be a key part of it. In other words, you are working on the biggest issue for the car industry currently. In the case of autonomous driving, we are even talking about entering into the history books. 

MICHAEL WÜRTENBERGER
It is true we are living in challenging times. But the fact that we are leading the way and playing an essential part in creating something big, motivates BMW, my staff and me personally. What could be better than working on the future? Anyone can handle the past, but the future is tremendously motivating.