An interview with Professor Dr. Sebastian Feldmann on automation, AI, and future collaboration between humans and machines
Despite the rapid pace of development, we probably only have a vague appreciation at the moment of the potential offered by systems based on artificial intelligence. What we do know for sure is that AI will bring about sweeping changes to the world of work. TRANSFER magazine talked to Professor Dr. Sebastian Feldmann, Steinbeis Entrepreneur and Dean of Studies at Aalen University, about opportunities and challenges. During the interview, the Steinbeis expert received some unexpected help answering our questions. Let’s start at the beginning.
Hello Professor Feldmann. Artificial intelligence and automation in general are shaping the future of the economy. SMEs often have limited financial and human resources, so for them introducing this form of technology often takes longer than for large companies. How can SMEs benefit from developments in this area and improve their competitiveness?
I believe the big issue here is accessibility. AI makes it possible to create new synergies, especially by offering barrier-free applications and interacting with their human counterparts. AI systems are capable of learning and improving themselves. Examples like ChatGPT are an illustration of this potential. In my opinion, SMEs will derive particular benefits from this. Employees can have more unnecessary and time-consuming tasks taken off them. In addition to that, for small and medium-sized businesses AI offers help when it comes to accessing new employee resources. So for example, AI in the form of a virtual conversation partner can act as an assistant in supporting older employees, or staff with language barriers, and help with certain steps within processes. Recent recruits, or employees from a migratory background, can use this support to find their way much more quickly around the company. These systems are multilingual and can share information beyond language barriers. Tasks carried out at reception or interactions with customers can also be made more automatic by using virtual assistants. This frees up more resources in the future for staff to work on creative processes and new developments.
I think using artificial intelligence will improve the efficiency of SMEs in the long term. But to tap into this potential, everything depends on the understanding people have at the company when it comes to using AI systems. Not only does it require a lot of ongoing training, but there’s an onus on senior management to ensure people feel enthusiastic about this technology.
What role does the transfer of know-how and technology from research into industry play in this?
In the future, systems based on artificial intelligence will serve as an accelerator for the development of products and creation-creation value-creation processes. This will pave the way for completely new categories of products, but there will also be sustained enrichments in existing products and processes.
The computational tasks carried out by AI often happen in the cloud, which offers good connectivity to other systems and the facility to link up with a whole variety of enterprise applications. There will be no more instances of AI systems only being used for individual purposes. There’s a clear trend toward developing general-purpose AI systems that generate context by incorporating feedback loops from different data sources.
The availability and coherence of information supplied by sensor data, or user interactions, will play an increasingly decisive role in this. There will be a fundamental change in how new knowledge is generated and how research is transferred. For example, how will AI developed through research be introduced to corporate processes in the future? It can no longer be the case that ‘only’ one mathematical formula will be supplied for calculating operational stability.
With AI applications, you’re often dealing with complex models that autonomously support certain decision-making processes and generate new knowledge. At the moment, sharing know-how is mostly a manual process involving highly specialized expert knowledge. I think this is where a new kind of process will enter the equation: The output of research groups looking at AI systems will be transferred into the industry. This will also be reflected in the training offered at universities. Maybe there’ll be degree programs for industrial AI artists or AI communication psychologists.
What sort of issues do your customers come to you with?
To display an example of a typical problem, let me describe a customer scenario of the previous year. A company contacted me and asked if artificial intelligence could help them to better understand errors in a production process and their causes. The customer application targets an extremely complex production process of a German Tier-1 supplier that is used to produce fuel rails for automotive engines. The piping system has to be extraordinarily fluid-tight due to quality demands. Thus, it has to be soldered together in a furnace. The AI that my team has developed is able to conduct cloud-based analysis of interdisciplinary data types of the production process, including robots, sensors, and control units. Due combination of various AI-based methods we were able to pinpoint the root cause of errors independently at the demo datasets. But it was found that data consistency and completeness are just as important as the model of the AI that had been trained. Some of the process information wasn’t even being captured by the sensors and control systems, or there was no way to capture it. So another system was needed to take into account real-time information supplied by people in production. For the development project this led to, we organized an AI workshop to focus on accessible information gathering. Another typical application in industry is direct collection and evaluation – during live processes – of information regarding quality levels. Such information is used to adjust machine parameters.
AI offers some amazing opportunities, but for many, it also leads to uncertainty and worries regarding job losses. Do those issues also come up in your projects?
Modern AI systems are already able to create art, tell stories, or even – independently under guidance – perform complex programming tasks. There’s going to be massive upheaval in the world of work and education, and this will make it necessary to fundamentally rethink existing job profiles. And it’ll be difficult to keep up with the pace of developments in this area. Maybe I can offer some food for thought on this: Computers were introduced on the premise that people could be relieved of simple and repetitive tasks, and thus have more time for the important things in life. Introducing robots to production was based on the same rationale. But in reality, exactly the opposite has happened: Because of the specifications used for machines, work packages have been made extremely simple, or highly complex. Tasks in the work environment have generally become extremely intensive. People can be reached everywhere and we urgently need specialists who are familiar with the increasingly complex nature of technical systems. In my opinion, there’s a shift on the horizon toward a different, collaborative way of working. AI will be perceived as a genuine virtual assistant and it will be supported and monitored by humans. But we need to be careful that “normal” employees do not become further detached from these developments.
Let’s be daring and gaze into the future of AI and automation. Where do you believe the journey is taking us, especially for SMEs?
“In the future, AI and automation will penetrate deeper and deeper into the SME sector. There will be more ways to automate processes, analyze data, and make predictions to help companies optimize and streamline their business operations. Continuing developments in machine learning and other AI technology will enable SMEs to make more effective use of resources and improve their competitiveness. But it will also become increasingly necessary to consider the ethical and social implications of technological developments, to ensure that the benefits of technology are distributed equitably and that human rights are respected.” Would you ever have suspected that this answer to your question didn’t come from me? I’ve just had a bit of help from AI and used ChatGPT, which a lot of people are talking about at the moment. The company, which Elon Musk is also a shareholder of, is a pioneer in making AI applications suitable for the masses by 2023.
This clearly displays, that in the future evolution of AI-based work scenarios, real interaction with different forms of artificial intelligence will not only be possible but almost inevitable. The texts it generates are indistinguishable from the responses of colleagues or customer service representatives. This opens up completely new realms of possibility to identify machine defects or interact with customers and employees. AI in the form of robots or virtual avatars will be perceived as actual interaction partners for humans, and it will help people navigate their way through the increasingly complex processes of companies.
Prof. Dr. Sebastian Feldmann (interviewee)
Steinbeis Transfer Center Mechatronics and Robotics (NectOne) (Aalen)