“We only see AI as a new tool. A very powerful tool, admittedly.”
AI might not have come up with the idea of generating electricity from solar energy. But you would only have needed someone with a flash of inspiration to do so. The idea could then possibly be implemented by AI, which would put electrical engineers out of work. Right?
No, it certainly wouldn’t put them out of work. But it will support us in the development and product creation process. And that’s a good thing.
People assume that AI is just an additional tool and that such technical progress – as was the case for the industrial revolution or digitalization in the past – ultimately always has a positive impact. However, is AI really “just” a tool, or is it more than that?
We too have discussed this issue here among colleagues. And yes, we do indeed see AI as just a new tool. Admittedly, it’s a very powerful tool. A tool that can do more than the ones we had in the past. But for us, it still remains a tool that primarily helps us in our work. Take programming, for example. Of course, I could also do it on my own and work through all the manuals myself. But I would be much slower than AI. All in all, this means that I currently view AI as nothing more than a very good new tool.
As you just said: “currently”! But anyone who decides to study electrical engineering today still has 40, perhaps 50, years of professional life ahead of them. And it is impossible to predict what artificial intelligence will be able to achieve in that time.
It is true that progress in this area is proceeding at a rapid pace. I myself have just tested how ChatGPT has developed over the last year and a half alone, and the improvement is tremendous. It is therefore difficult to predict what will happen in 20, 30 or even 50 years’ time.
Nevertheless, I am very confident in the long term, especially in the case of electrical engineering and information technology. After all, our discipline stands above all for products, be it fuel cells, energy or data storage systems, to name but a few. Our discipline is so versatile and is needed everywhere, so I’m really not at all worried that we might no longer be needed in the distant future – though of course I don’t know exactly what the job of an electrical engineer will look like in 50 years’ time.
You have just emphasized how quickly AI is developing. But is it even possible for us humans to keep up with the development – especially for people who work with it, including electrical engineers?
I am well aware of the problem. I myself am not an AI person by nature. But you have to make an effort to stay up to date and acquire the relevant knowledge so that you can join in the discussion and have a sound basic vocabulary. Of course, you won’t have the know-how to understand every single neural network straight away, but you will have a certain basic knowledge that will enable you to familiarize yourself with everything that is required. Basically, none of this is anything new for us. As an engineer, you always have to stay on the ball technically anyway and, above all, learn the appropriate methods in order to be prepared for all requirements. I would even go so far as to say that it might actually become easier for us. After all, AI also provides us with a new learning tool to teach us new things more effectively.
“We want to enable our students to apply AI in practice.”
Then let’s proceed to look at how the electrical engineering degree program needs to change and what you are doing differently in your degree program in Karlsruhe.
We have been offering the new “Artificial Intelligence in Engineering” course for a year now. On the one hand, it covers basic engineering knowledge, which is not fundamentally different from courses in mechanical and electrical engineering. On the other hand, we combine this with the topic of AI, i.e. methods of AI, implementation variants of AI, areas of application of AI and the verification of AI. One reason why we do this is that it goes down well with young people and, like other universities, we are of course also looking for new students. However, we are also doing it because we see it as the future and want to bring students into contact with the topic at an early stage and enable them to apply AI in practice.
And it may also show young people that they don’t have to study computer science to work with AI. Right?
Electrical engineering, mechanical engineering and computer science are all becoming increasingly intertwined anyway. Our work perhaps revolves somewhat more around the hardware, such as the control unit, the cell phone or the microcontroller. Computer scientists may not be quite so familiar with such devices and that’s where we as engineers can perhaps make more use of our expertise.
Incidentally, I also asked AI itself, in this case ChatGPT, whether AI posed a threat to the electrical engineering labor market. It too said that the introduction of AI would lead to a change in the job profile rather than a massive loss of jobs.
Look for yourself! And I’d like to present one more argument that backs this up. I simply believe that one of the characteristics that sets us humans apart is the fact that we are social beings and work things out as a team. When I think back to my earlier days at Bosch, to how we developed products together with VW and other customers, I just can’t imagine that a machine would be able to implement the necessary interaction independently.
This interview was not only transcribed with the help of AI, but also shortened to the required length. It didn’t work perfectly, but the result was not bad at all. It was only necessary to intervene in a few places, for example because the follow-up question would not have worked otherwise. Incidentally, the questions were still asked by a person.