It is a widespread notion: At some point, machines will completely take over human work. “We should question this image of the future,” says Dr Norbert Huchler from the Institute for Social Science Research Munich (ISF). “It blocks development opportunities if we think of technical progress in industrial production only in terms of full automation. Rarely do we consider how automation and human labour complement each other in new ways. How can humans and machines work together – and in such a way that employees develop in the process? How can we make work and technology in a factory humane, conducive to learning and motivating?”
The sociologist of work, together with an interdisciplinary team, is shedding light on the possibilities of “Empowerment in the Production of Tomorrow” (EmPReSs for short) as part of a project funded by the bidt. The researchers want to rethink mixed skill factories and collaborative robot systems.
The bidt project focuses on two questions:
- How can the increasingly complex interaction between humans and robots be organised as precisely and efficiently as possible?
- How can this be done in such a way that it contributes to the empowerment of employees and the humanisation of work in production?
Important terms briefly explained
Parts of the work process are transferred from humans to artificial systems (machines, robots, artificial intelligence). The aim is to increase efficiency. Jobs can be lost; others are created.
Empowerment or enabling; in the context of the bidt project EmPReSs, it is about work design in industrial production that is conducive to experience and learning.
Mixed Skill Factory
Humans and machines work together in a mixed skill factory. Their different skills are flexibly combined. This makes work easier for employees and more interesting. Human-machine interaction helps to try out new things and learn skills.
If a robot works as a semi-autonomous system, e.g. in the production process, together with humans and is not separated from them, e.g. by a protective device, it is called collaborative.
Practical example: Tricky cables
To explore the opportunities and risks of mixed skill factories and collaborative robot systems, the researchers are investigating various use cases. One of these empirical cases is the wiring of a control cabinet: “This houses connected electrotechnical parts to control a large machine, for example,” explains Florian Lay, PhD student at the Institute of Robotics and Mechatronics of the German Aerospace Center (DLR). “Wiring is not easy, because even within the same production series there are minimal deviations and many cables laid on top of each other behave unpredictably.” For a robot, this is a tricky task that has not yet been solved.
In an industrial operation, there are numerous activities of this kind that no machine can completely take over. Here, it depends on the interaction between humans and robots. In the case of the control cabinet, the only question was: How far can the wiring be automated? How far can the robot get on its own? If there are problems, the human has to intervene or continue working, so it’s not a truly collaborative solution.
The team of the bidt project therefore proposes to systematically consider the interaction between humans and robots from the very beginning and to design it interactively. Empowerment in this human-machine interaction presupposes that not every process is automatically controlled technically. Employees must be able to decide for themselves to a certain extent when and how to use the robot.
Mixed skill factories as a driver of success
Thanks to years of experience, employees are often the best judge of how robotic systems are sensibly designed and used in production. “It is important to get the stakeholders on board as early as possible when work systems change or are newly initiated,” explains Norbert Huchler. “Automation is supposed to make production more efficient. If workers are left out of the planning, inappropriate systems may be purchased or employees may not accept the new ‘tools’.”
Automation is not a one-way street, we have to think it together with work. This is only possible if workers are involved from the beginning.Dr. Norbert Huchler To the profile
In the worst case, the robots just make extra work. Alexander Perzylo from the Bavarian State Research Institute for Software-Intensive Systems (fortiss) reports: “We have seen empirical cases where only one person in the company can operate the robot. If that person is sick or on holiday, the expensive technology sits in the corner and everything is done by hand.”
Can artificial intelligence help with work organisation?
In addition to the robotic system, the project team is investigating the possibility of using artificial intelligence (AI) to organise work. Florian Lay explains, “We want to show how AI as an assistance system can help to better manage work in the background, for example by assigning tasks in the so-called ‘mixed-skill zone’ based on humanising and learning criteria – that is, in the area where both humans and machines can process tasks.” The AI recognises and indicates how humans and robotic systems best complement each other in a given task.
But how does the AI know who does what best; who does a task faster and more accurately; who has what skills or functions? To prototype this, the research team captured all process step requirements of an empirical case in a formal representation. Only when all the necessary skills are precisely described can the AI sensibly distribute manual tasks to humans or robots and optimally combine their abilities.
“It was a great challenge to generate the necessary data basis for the AI system,” says Alexander Perzylo. “There are numerous sources of information that describe what a robot can do: for example, grasp a component and place it in another location or screw in a screw. But for tasks that humans do, these detailed descriptions don’t exist.” You can imagine it: You have to explain even simple hand movements precisely to a robotic system. An expert looks at a plan and knows what to do. She has experience and uses her skills as well as flexibility and intuition. The project team had to go to great lengths to create this knowledge as a basis for the AI.
“This is a great added value of the project,” emphasises Norbert Huchler. “We see quite clearly that you cannot simply automate a process without changing its conditions. The robot just doesn’t do it exactly like a human. Appropriate framework conditions would have to be created first.” The AI must take this into account when distributing tasks. In the best case, it also assesses whether a task is interesting and conducive to the employee’s development or whether it makes ergonomic sense.
Work-immanent learning in the “factory of the future
With technical advancements, machines in industry will perform more and more work in the future, especially routine tasks. This prospect stirs up fears and reservations among employees. Will the job on the production floor disappear in the future? Will human labour be replaced by machines? One thing is certain: job profiles and the required qualifications in an industrial company will change, not least due to interactive AI systems and collaborative robots. However, this does not have to be a risk for employees, but rather offers numerous opportunities for the working world of tomorrow if the transformation of work is thought through from the beginning and designed in an empowering way.
“The complexity of human work is increasing. It is crucial that employees are constantly learning how to deal with technology,” explains Norbert Huchler. This is the essential requirement for mixed skill factories and their employees: to enable, accept and implement learning immanent to the work. In this way, people with less technical qualifications would also benefit from the Mixed Skill Factory, explains Florian Lay. “In the project, we are developing an AI-based robotic system. With it, we want to show how all employees are enabled to acquire skills while working and to take on complex tasks in cooperation with the machine.”
We are developing a system that shows how human-robot collaborations can be designed to be efficient, experiential and conducive to learning.Florian Lay To the profile
“In doing so, we don’t just consider learning as imparting knowledge,” adds Norbert Huchler. “The AI should provide contextual information so that employees understand: Why am I doing this right now? But it should also open up room for manoeuvre within the human-machine interaction, so that people can try something new, change a procedure, design and experiment themselves in order to build up experiential knowledge and skills.”
Collaboration 4.0 – humans are not replaceable
“Our advantage at EmPReSs is the interdisciplinary view, which opens up new perspectives on human-machine interactions,” says Florian Lay. “In robotics, people rarely ask which social aspects should be included in the design of technology or how humans and robots, with their different potentials, can optimally work together. The focus is usually on technical feasibility.” In the bidt project, researchers from the social sciences and technology development are jointly investigating new concepts for the collaboration of humans and robots in factories, because Industry 4.0, the digitalisation of industrial production, is not a purely technical project for the future. It is dependent on people.
The results of the bidt project provide material for further research questions and are enormously relevant to practice. “Every automation step brings new tasks that only humans can handle, whether because of the necessary dexterity or intuition – you can hardly implement that in robotic systems,” says Alexander Perzylo. “Innovative human-robot collaborations have a decisive influence on the economic success of industrial companies.”
Automation makes value creation more complex. Production will continue to rely on people with their special skills in the future.Alexander Perzylo To the profile
Production without humans? Doesn’t work. That is why it is essential to make work processes in industry more flexible through “hybridity” (human and technology) and to include the needs of employees in the process. Innovative models of the “factory of the future” could also provide new orientation on how the interaction between man and machine can be reinvented again and again. The bidt project therefore raises the socially relevant question: Can we organise everything through technology or do we want to design production with hybrid systems in such a way that people maintain and develop their skills in interaction?