Hallo, Maschine! Wie geht es dir?

Chatbots mit künstlicher Intelligenz (KI) können die Arbeit der Programmierung und der Diagnose von Automatisierungskomponenten revolutionieren.

Siemens und Microsoft entwickeln Industrial Copilot

Die Revolution wird durch den Industrial Copilot eingeleitet, den Siemens mit Hilfe von Microsoft entwickelt hat. Der Copilot schreibt den Programmiercode für Maschinen in Fabriken und hilft, Softwarefehler aufzuspüren. Die Anwendung wird jetzt in der Praxis getestet – und könnte später über Siemens Xcelerator angeboten werden.

ChatGPT war von Anfang an ein riesiger Erfolg. Seit OpenAI im November 2022 den Schleier über dem Chatbot gelüftet hat, vergeht kaum ein Tag, an dem die Nutzer*innen nicht erstaunliche und nützliche Dinge mit ChatGPT tun.

Der Chatbot wird jetzt zu einer Arbeitskraft in der Fabrik. Mithilfe der natürlichen Sprache können menschliche Mitarbeiter*innen in einer Fabrik mit Maschinen chatten oder Automatisierungscodes schreiben. Diese Fähigkeit nimmt eine enorme Last von den Schultern der Spezialist*innen für Fabrik- und Prozessautomatisierung, die sich mit der Programmierung von speicherprogrammierbaren Steuerungen auskennen.

Und das alles kommt genau zum richtigen Zeitpunkt. Der Markt für diese Fachkräfte ist leergefegt. Hinzu kommt, dass die geburtenstarken Jahrgänge der Babyboomer aufhören und in den Ruhestand gehen und ihre jahrzehntelange Erfahrung mitnehmen.

Code auf Knopfdruck

Ein Siemens Industrial Copilot, der Werbung programmiert, in der ein Roboter und eine Person gemeinsam an einem Projekt arbeiten.

Siemens is feeling this pressure, too. The world market leader can sell its automation technology only if the customers have programmers who can program and operate it.

For this reason, it is hardly surprising that Siemens has taken a deep dive into artificial intelligence and ChatGPT. This chatbot has quite a skills set, from composing newspaper articles and poems to speaking a few programming languages. The code is so good that it can save human programmers a lot of time and minimize bugs — especially in automation technology.

The trouble with automation code is its vast array of repetitions. The job of programming it consumes tremendous amounts of time. But the highest costs of all are frequently caused by minute bugs in the code. These bugs can never be completed eliminated and can be the source of painful disruptions in plant operations.

Enter an AI-based digital assistant that can, not only replicate an automation code as many times as necessary in a matter of seconds, but can also minimize bugs.

For this reason, it simply made good sense for Siemens, the world's largest maker of automation technology, and Microsoft, the world's largest software company and the most vocal booster of OpenAI, to work together on joint solutions. To do so on the highest level, Microsoft CEO Satya Nadella and Siemens CEO Roland Busch never miss an opportunity to boast about the synergies being generated by the two companies' partnership.

Chatting about searches for bugs

Siemens is developing and testing its Industrial Copilot, a technology that helps automation engineers do their jobs and could be offered to users, for example, on Siemens Xcelerator.

The open platform assists industrial customers with their digital transformation — offering a selected portfolio of networked hardware and software, a high-performance ecosystem of partners and a comprehensive marketplace of applications.

The Industrial Copilot is currently being employed in two use cases.

First, it is helping write PLC programming code. This code lets Siemens control systems know what the machines have to do in which order and under which conditions. In the process, the Industrial Copilot does something more than generate PLC code. Industrial Copilot can also translate PLC code from a different programming language if necessary and optimize it.

The second area is troubleshooting when a machine breaks down. The copilot not only locates the bug, but also suggests ways to exterminate it. This alone is something astonishing.

But what is really fascinating is the way that maintenance employees do their jobs. They chat with Industrial Copilot, explain the problem, ask for details and discuss suggested changes — all in a natural-written language like with ChatGPT or the Bing chatbot — as though they were talking to a colleague.

Titled, 'Siemens Industrial Copilot: AI Copilot supports troubleshooting,' this illustration shows a person and a robot working together.

Industrial Copilot brings IT and OT together

Siemens and Microsoft performed some initial demonstrations of the Industrial Copilot at Hannover Messe 2023. "The customers were excited," said Kai Brasche, Vice President of Global Strategic Alliances at Siemens Digital Industries.

Olivia Klose, Global Partner Technology Strategist at Microsoft, confirmed the great interest of the visitors: "Together, we're building the bridge between IT and OT. It is all just like the bridge between Microsoft Azure and Siemens Xcelerator."

While Siemens provides the automation part of the Industrial Copilot and can later make it available on the Siemens Xcelerator platform, the natural language processing is carried out via ChatGPT using the Azure OpenAI Service in Microsoft's Azure Cloud. This creates optimal security, data protection and reliability.

Customers see none of this interaction — to them, the Industrial Copilot is simply an all-encompassing service.

Initial pilot projects at Siemens plants

Pilot project customers have already been found, and, as is often the case at Siemens, they come from within the company. The plants in Amberg and Erlangen, which produce automation components, are highly automated and use Siemens technology to get the job done.

The team led by Erik Scepanski and Claus Romanowsky is going through some use cases and sizing up the Industrial Copilot's strengths.

"We are also assessing whether the Industrial Copilot can be used to implement diagnosis and troubleshooting in a smartphone app or in the user interface of a machine," the innovation manager Scepanski says.

"If a machine breaks down on the weekend, the technician responsible could consult with the chatbot and initiate a solution — from the comfort of his or her own couch at home," Romanowsky says.

"The Industrial Copilot will revolutionize factory buildings," Kai Brasche predicts. In growing industries like battery production, it will significantly reduce the required time to market introduction and enable manufacturing to be more flexible. "And this is just one of many fields of application."