Photo from rawpixel.com
Prof. Dr. Shadan Sadeghian
How does the world of work change when intelligent systems take on the role of employees? How should they be designed so that team members experience working with them as meaningful? And where do conflicts typically arise? These are the kinds of questions that Junior Professor Dr. Shadan Sadeghian is investigating at the University of Siegen. She is a junior professor for Autonomous Interactive Systems at Schools III and researches the interactions between humans and artificial intelligence.
A robot arm twitches briefly when it starts work, but it doesn't say "Good morning". A chatbot solves a problem with a customer, but it doesn't drink coffee during the break. And software automatically reads in a bank transfer, but can't do anything with an after-work beer. Other systems can recognize early on when important parts for production are missing and order them in good time. They work precisely and quickly, sometimes know more and see connections better than their human colleagues. Collaboration is automated - but not automatically free of conflict.
There can be many advantages when automated systems accompany work or take over tasks. "Automation has meant that people no longer have to do many routine tasks," says Sadeghian. On the assembly line, for example, machines have been doing repetitive and monotonous work for decades. Now, however, the tasks that machines can perform go far beyond routine activities: systems with artificial intelligence (AI) can process information, derive patterns from it and act independently - for example, placing orders.
In many areas, it could sooner or later come down to people working not only with human colleagues, but also with robots. This will change the role of people at work - and therefore possibly also how satisfied they are with their work and how meaningful they perceive it to be. The scientist often hears concerns that AI could take jobs away from people. "However, I believe that it will be more about the extent to which tasks change through the use of AI."
She investigated such effects together with Professor Dr. Marc Hassenzahl from the University of Siegen in the experiment "The Artificial Colleague". A total of 104 people who work with knowledge, such as programmers and scientists, were interviewed in an online survey. They were asked to evaluate a fictional scenario involving collaboration with an AI. "Many workers experience collaboration with people as more motivating and meaningful than with artificial intelligence (AI)," says the researcher. "This can be difficult because such technologies have the potential to fundamentally change work."
Specifically, the experiment involved organizing a meeting and organizing the time and space as well as preparing the content for the meeting. An AI-controlled system took over some of the tasks for some of the interviewees. The respondents were then asked to assess which tasks they found important and meaningful and how satisfied they were with the outcome of the meeting. "This showed that it was important for the majority of respondents to take on the tasks that they perceived as important and meaningful," says the researcher.
In a second online experiment, Sadeghian investigated the distribution of tasks between humans and AI with regard to job satisfaction in shift planning for nurses in a hospital. "There is a trend to treat AI as an instrument that we supervise and whose results we control," she says. "But the research shows that people find their work more meaningful when an AI is their teammate." In this case, they feel more responsible for the results and are more satisfied.
In most cases, however, tasks in companies are not distributed according to these criteria; instead, intelligent and automated systems are introduced in order to increase a company's performance and thus its profits, according to Sadeghian. Employees often have no say and have to come to terms with the situation. What counts for a company is the result of the work. "However, how satisfied people are with their work does play a role in success," says Sadeghian. People who are satisfied and find their work meaningful are less likely to change jobs, are more motivated and are more committed. This is another reason why the scientist is looking into the question of how collaboration between human employees and AI systems should be designed so that people continue to find it satisfying and meaningful.
Satisfaction in the human-AI team
Whether someone is satisfied with their work or not depends on various factors. "These include a sense of achievement as well as a good relationship with colleagues or being proud of one's own achievements and abilities," says Sadeghian. The study also showed, for example, that it plays a role whether someone feels they have control over their work and the results. "However, we don't know what it means to be proud of the joint result in a human-AI team if the AI can't be proud."
More and more potential fields of application are currently opening up for AI systems: they can not only be useful for creating shift schedules in hospitals or factories, for example. "They are also likely to play an increasingly important role in other areas, such as journalists and designers," says Sadeghian. People will probably continue to play an important role in producing results. But this could also change the tasks considerably: Those who used to write journalistic texts, for example, could in future check texts for accuracy that have been created by an AI. "The key question is: if tasks change in this way, will people still be satisfied with their work?" asks Sadeghian. It can become problematic, for example, if an AI takes over the part of the work that someone likes the most because it seems to be better equipped for it.
"It is always important who takes on which part of the tasks so that people also have the feeling that they have contributed to the result in a meaningful way," says the researcher. Working with AI does not necessarily have to lead to problems: If it is used sensibly, it can even increase job satisfaction. "It is conceivable, for example, that someone acquires special skills to operate the AI," says Sadeghian. "This experience can be very meaningful."
This text first appeared in the Future research magazine of the University of Siegen:
When an artificial intelligence becomes a team member (Maria Berentzen), page 39