All right – so far, so good! But why are the researchers interested in how competent or trustworthy a robot that speaks the Berlin dialect appears? For cognitive psychologist Katharina Kühne, the study indicates the situations in which it might make sense for a robot to speak a dialect. After all, robots will be increasingly present in our everyday lives. For example, a robot could use dialect to communicate with elderly people in need of care, while a sales robot in a restaurant might speak in standard language. “Competence is more important in sales consultations, trust in a nursing home,” she summarizes. The article on the study was published in early 2024 in the journal Frontiers in Robotics and AI. The media response was enormous. And six months later, Kühne still gets visits from journalists.
Fascinated by artificial agents
The scientist holds a master’s degree in linguistics and cognitive psychology and is now pursuing her doctorate under Martin Fischer, a Professor of Cognitive Sciences at the University of Potsdam, who has been researching human-robot relationships for years. “Martin Fischer gave me the bug,” Kühne says with a smile. After graduating, she couldn’t let go of her fascination for the effect of artificial agents on humans. What role does human resemblance play when we interact with them? What does an anthropomorphic design do, what do human-like abilities trigger in us? “When an agent moves, this is enough for us to perceive him as a social counterpart. We then attribute emotions, intentions, and thoughts to it,” Kühne says. She conducted several online studies, starting with the voice. Is a humanoid voice more popular than one that sounds more technical? “This question is also intriguing because we know that artificial agents may be perceived as scary if they look very human-like.” This is what the theory of the “Uncanny Valley” says. “In our study, however, we found that the ‘uncanny valley’ does not exist in the voice. The more human-like a voice is, the better.” Another study examined the general perception of robots. Do we experience them as helpers or even friends? “During the Corona pandemic, many people suffered from loneliness. We found that loneliness increased acceptance of a robot at home.”
“Hello Nao! How are you doing?” Kühne says to Nao and asks him to wave and sing. But he is on strike. The scientist gently strokes its head. Although we know that Nao has no feelings, we say things like, “He’s having a bad day.” or “He’s grumpy.” when he doesn’t follow our commands. Kühne has an explanation for this: “Whenever an artifact sends out social signals, it triggers something in us. That’s in our nature.” In interaction with robots and other artificial agents, we adults also react like children at play. We are in an ‘as if’ state, doing things against our better judgment.
However, it is difficult to be angry with Nao when it does not do its job. The robot just looks too cute with his short arms akimbo looking at us with big eyes. If we had a two-meter tall robot in front of us, we might be scared, Kühne says. The human-like androids Erica and Sophia or Geminoid HI-1, the “twin” of its creator Hiroshi Ishiguro, may also seem scary because they look like real people. Nao, on the other hand, has a “medium human resemblance”. In principle, however, it takes almost nothing to humanize an artefact. Kühne reports on experiments in which scientists were able to show that movements of geometric shapes such as triangles and squares on a screen are enough for people to imagine social interactions, even entire stories that unfold between the abstract shapes.
Do we want robots with consciousness?
In Japan, people already encounter service robots everywhere. They take orders in restaurants and bring food to the table. „Because they resemble humans, people often speak to them. But as a rule, they can’t just start chatting.“ The researcher calls such one-sided interactions “parasocial”. Even in German nursing homes, where real pets are not allowed, companion robots in the shape of animals are becoming increasingly common. “Of course, this raises ethical questions: Do people with dementia understand whether they are petting a real cat or a robot designed to resemble a soft, cute cat? Or does it not matter as long as it benefits them?” The linguist asks herself the same question when she participates in robot fan communities on social media. “Many people buy their companions dresses and hats. Some have three or four robots at home, and thus a lot of action. They also take them with them when they travel.” Of course, one could ask whether it would not be better to develop relationships with real people. But where this is not possible, robots may be a helpful alternative.
Although the researcher treats Nao almost tenderly, she doesn’t feel guilty when she locks it in the closet in the evening. “I turn Nao off after work and that’s it. We still have control over the artificial agents, and that’s a good thing.” But what if Nao were to review the day in the closet at night or feel lonely? “The fact that robots develop consciousness is both an incentive and a fear. Would we understand their language, would we even perceive it? After all, we don’t know what it’s like to be a robot – our experience is only good enough for ‘human’ feelings.”
„Eenfach wunderschön, erstaunlichet Kunstwerk!“ (Simply beautiful. An amazing piece of art!), Nao says in the Berlin dialect at the end of the video about Picasso’s painting. That is something you could also say about the robot and its peers.
In addition to Katharina Kühne, other authors of the Online Study include Erika Herbold, Tristan Kornher, Dr. Yuefang Zhou, and Prof. Martin H. Fischer (University of Potsdam) as well as Prof. Oliver Bendel (University of Applied Sciences and Arts Northwestern Switzerland).
Article
https://www.frontiersin.org/journals/robotics-and-ai/articles/10.3389/frobt.2023.1241519/full
Katharina Kühne studied linguistics at Freie Universität Berlin and cognitive psychology at the University of Potsdam. She then completed further training as a clinical neuropsychologist. This field focuses on the diagnostics and rehabilitation of patients with organic brain injuries.
This text (in german language) was published in the university magazine Portal - Zwei 2024 „Europa“ (PDF).