Hanover The tablet that Hewlett Packard Enterprise (HPE) placed on its stand at the Hanover Fair helps the user to operate a robot – and can replace a service technician. HPE and the Heidelberg start-up Aleph Alpha are demonstrating an assistant at the trade fair on Monday that uses artificial intelligence (AI) to communicate with factory staff using natural language and images.
This involves questions relating to the installation, maintenance and operational safety of the robot. The AI assistant acts like a service technician who supports factory personnel in solving very complex tasks.
“Because the factory staff does not have to read manuals that sometimes have several thousand pages, processes are accelerated,” explains HPE Germany boss Marc Fischer. “And in an emergency, the AI assistant can guide the employee to stop production processes.”
The AI-based application also makes a significant contribution to occupational safety, explains Fischer. In addition, the AI assistant always has the appropriate sources from the manuals ready for his answers. Aleph Alpha has been developing artificial intelligence on an HPE supercomputer since 2019, which competes with the AI projects of US companies such as OpenAI and Google.
In particular, the speech-based AI program ChatGPT from OpenAI has caused a stir in the past few months. The chatbot uses artificial intelligence to understand human language and, in many cases, generates linguistically convincing answers.
The topic of AI is also dominant at the Hanover Fair. Industrial companies hope to use AI to make their processes safer, more flexible and more efficient.
Industry managers attach great importance to AI
“ChatGPT also means a revolutionary step for the industry,” says Michael May, who heads the technology field of data analysis and artificial intelligence in Siemens research, the Handelsblatt. In the future, a service technician will be able to talk to a machine very easily – and the AI can, for example, help to write programs for the control.
For example, the manuals for Simatic, a programmable logic controller (PLC) from Siemens, have 12,000 pages. Nobody would have to go through these anymore if AI chatbots are used.
May is not alone in his opinion. According to a global survey of 1800 industry managers, nine out of ten companies surveyed consider AI to be indispensable in production. Almost two-thirds of managers primarily expect an increase in production efficiency from AI-based applications.
Indonesia’s President Joko Widodo (left) and Federal Chancellor Olaf Scholz (2nd from left) get information from Siemens boss Roland Busch.
(Photo: IMAGO/Die Videomanufaktur)
According to the survey by the Boston Consulting Group, the use of the technologies is also financially worthwhile: companies that successfully use AI have been able to reduce production costs in the areas of application by an average of 14 percent. In Germany, three quarters of the manufacturing companies surveyed have already fully rolled out AI in at least one application – the proportion is only slightly exceeded in India and China.
Read more about artificial intelligence
May predicts that AI will find its way into industrial practice faster than many expect. However, the models still have to be weaned from “hallucinating” – i.e. inventing facts – and they still have to work with Siemens for use in the Siemens cosmos, for example – and customer-specific knowledge are fed. But that will take months rather than years. “It’s our turn to work on such applications.” Several hundred AI experts work in research in May’s team at Siemens, and the group has a total of over 1,400 employees.
For the foreseeable future, however, artificial intelligence will only ever be an auxiliary assistant, says May. “We should concentrate on building systems that enhance human intelligence, and not ones that are fully autonomous.”
Language models like ChatGPT could be of great help, especially in the industrial metaverse. “The user can use it to conduct dialogues and research.” This can help to find errors and optimize processes. Simulation and design development could also soon benefit from the use of artificial intelligence.
Industry is already saving millions with AI
A year and a half ago, Schneider Electric decided to fill the role of Chief AI Officer, who basically deals with all internal group topics related to AI, explains Peter Weckesser, Chief Digital Officer of the company. At Schneider Electric, products are manufactured with AI-based software to optimize process flows – but also to optimize algorithms for saving energy. “On the other hand, we also use AI internally – for example to optimize and monitor our supply chains or for our customer centers,” says Weckesser.
This year, the company will save a three-digit million amount through the use of AI. “The customer inquiries alone that could be answered automatically by the AI offer enormous savings potential,” says Weckesser.
>> Read also: Bosch wants to bring AI into the products faster – and is working on solutions with ChatGPT
Schneider Electric already has a high demand for employees who can work with AI-based technology. “To this end, we offer AI upskilling programs for our employees – because everyone will come into contact with it at some point,” predicts the company’s digital expert.
“When robotics was being developed a few years ago, many were already afraid that jobs would be lost as a result of AI. The opposite has happened – AI creates jobs and at the same time helps against the shortage of skilled workers,” says Weckesser.
Read more about the Hanover Fair
Fero Labs, a young high-tech company with locations in Düsseldorf and New York, offers software solutions for the use of “machine learning” in the steel, chemical and cement industries. The AI-based application ensures that emissions in the production process are reduced and costs are reduced while quality remains the same.
Fero Labs customers include international corporations such as Henkel, Volvo Trucks, Gerdau, Evonik and Covestro.
“The program uses algorithms to calculate the parameters under which production must be carried out, for example to guarantee a specific elasticity, tensile strength or yield point for the steel while it is still liquid,” explains Tim Eschert, Managing Director of Fero Labs. The application collects data from the past six months and checks which parameters lead to the desired quality.
A special feature: “Our models, which we use in industrial production, say in addition to their analysis, whether they are sure about it and how they came up with it,” explains Eschert. The evidence is about numbers or statistical relationships. “White box machine learning means that the model makes a statement and also communicates the area of uncertainty around it.”
Engineers must become digital engineers
Artiminds also relied on the integration of artificial intelligence early on. The Karlsruhe company is developing software with which the use of robots can be planned and programmed. “If the robot is supposed to feel or see like a human, the requirements are high,” says co-founder Sven Schmidt-Rohr.
With the help of the software, the robots can decide for themselves how much to accelerate or how tightly to grip when they are grinding something or grabbing irregularly shaped objects, for example. The system is constantly learning and optimizing itself during operation, says Schmidt-Rohr.
There are now also AI language models such as ChatGPT. “With deep learning, we have laid an industrial foundation that we can now dock onto,” says Schmidt-Rohr. The system can then, for example, better decide which step to take first.
In addition, the programming of the robots can be significantly simplified with the help of the models because a dialogue is possible.
“Then I can say, for example, that I want to grind something. The query could then be what material it is about.” Such programming via dialogue could be possible in one to two years.
According to Schmidt-Rohr, the increased use of artificial intelligence will be urgently needed in Germany in the next few years. “The emerging shortage of engineers is a disaster for the location.” Medium-sized companies in particular should be more open to the new technologies. “In industry there is a lack of openness and the willingness to take risks to try out innovations quickly.” A change in mentality is necessary. “Engineers must become digital engineers.”
Siemens and Microsoft announce cooperation
However, no group will be able to dominate the topic of artificial intelligence in industry alone. And so new alliances are formed. Siemens and Microsoft, for example, announced a cooperation at the Hanover Fair. The two groups link the Siemens software team center for the so-called product lifecycle management (PLM) with Microsoft’s communication platform Teams and the language models of Azure OpenAI Service.
“Powerful, advanced artificial intelligence is becoming one of the most important technologies for digital transformation,” said Siemens CEO Cedrik Neike.
Specifically, the cooperation in the factories should look like this: With a Teamcenter app for Microsoft teams, experts in development, production and service should solve problems faster together. Using mobile devices, they can document and report potential design or quality issues using natural language.
The US tech group is cooperating with Siemens on artificial intelligence.
(Photo: Reuters)
Via Azure OpenAI Service, the app can then analyze this informal language data, automatically generate a summary report from it and then forward it to the design, development or manufacturing experts within Teamcenter.
In the future, digital assistants will also help to create and improve software codes for factory automation. Errors should then be detected more quickly during operation. To do this, Microsoft Azure Machine Learning and Siemens’ Industrial Edge analyze images and videos recorded by cameras.
“The integration of AI into technology platforms will profoundly transform how companies work and operate,” said Microsoft executive Scott Guthrie. Together, they unlock the power of AI for industrial companies, enabling them to “simplify workflows, break down silos, and collaborate more inclusively to accelerate customer-centric innovation.”
More: “De-risking instead of decoupling”: Scholz is pushing for a free trade agreement with Indonesia