The opening keynote entitled "Can machines build machines?" was given by Jan Seyler, AI expert at Festo. He also provided the answer: "Of course machines can build machines," Seyler clarifies. "When a robot assembles a car, that's exactly what it is: a machine building another machine". But the key question is: "Can AI build machines?" This does not primarily refer to the physical assembly of machines, but rather to the entire development process of a machine.

The traditional process of machine construction goes something like this: It starts with the question of what the machine is supposed to do, for example, screwing a screw into a thread. Mechanical engineers design the machine on a computer, select components, and plan the assembly to create a machine that solves this specific problem. Then, the machine is built, programmed, and put into operation.

From design and development to manufacturing and programming – up to now, humans have played a central role in each of these phases. "Of course, they have many digital tools at their disposal," says Seyler. But these tools are passive. "What if AI actively makes suggestions on what the machines could look like?" Seyler believes that algorithms could help optimize the machine design process.

How could such an AI-assisted development process look like? Here, too, it begins with defining the task: "Dear AI, design a machine that screws a screw into a thread." To accomplish this, the AI must have detailed information on the many thousands of parts available for building machines. This could include, for example, joints and segments of a robotic arm, or drive systems that enable the movement of the robotic arm. Seyler: "The AI must not only understand the capabilities of each individual part, but also know what new capabilities arise when two parts are combined."

To do this, the AI needs a precise digital description of all the available parts. The AI should also be able to use simulation models to test possible applications. "There are still challenges ahead. However, I believe we are very close to solving this problem," says Seyler. "In the near future, AI will design machines that actually work in the real world."

What does this mean for the profession of an engineer? Seyler: "The job description will have to adapt. Engineers will need to develop more creative ideas again, which they then implement with AI. The mathematical solution to a problem will be less important. Today, we talk a lot about the science of engineering. What we will soon see is the renaissance of the art of engineering."