On September 11, 2001, during the terrorist attacks in New York, the commander of one of the firefighting teams had an idea that he had to get his team out of the skyscraper they were in. He could not explain why he suddenly ordered everyone out.
In the book Thinking, Fast and Slow, psychologist and researcher Daniel Kahneman used the episode as an example of an intuitive decision: decisions we make quickly, based on a unique mixture of knowledge, experiences, and emotions.
Only later did the firefighter realize that he had reacted to the quietness in the room. In addition, his ears were unusually cold. It turned out that the core of the fire had been on the floor below. Just after the fire team escaped, the floor collapsed. The firefighter’s intuition saved the lives of everyone on the team.
– Most human decisions are intuitive. For now, robots are not able to do anything similar, Marieke van Otterdijk says. She is a researcher at the RITMO Centre for Interdisciplinary Research on Rhythm, Time and Movement at the University of Oslo.
Photo: UiO
The magic of human intuition
She does not know if a robot will ever be able to do what the firefighter in New York did.
Today robots can use artificial intelligence (AI) to make autonomous decisions; however, this decision-making is quite rigid and based on algorithms developed to find logical solutions.
Human intuitive decision-making is faster, more fluid, and based on emotions and associative thinking. For example, if we see a new person smiling and getting a gift, we might associate it with a birthday.
– This is not necessarily correct, which makes intuition tricky. However, intuition gives humans an adaptive ability. We can adapt to the changing world around us and even to robots. If robots get some sort of intuition, this can make them more adaptable as well. I do believe that it would increase their ability to help humans.
Intuitive robots in the making
In her PhD, van Otterdijk developed a theoretical framework that can form the basis for developing robotic intuition. The field is unexplored. It is also difficult, partly because we still do not know everything about human intuition.
Part of her research has revolved around studying human interaction with robots and the role of human intuition. Most architectures and algorithms built for robots take inspiration from, for example, psychological theory.
– It is essential to see how these theories play out in real interactions. Which aspects are most important, and which are potentially missing?
When she brought a robot to a nursing home, one of the first things the robot did was introduce itself and shake hands.
– We wanted to check how people responded. Many seemed to understand what the robot wanted to do, but some were unsure and looked at the researcher. After all, the robot was unknown to them. Some were also afraid of their hand getting hurt.
In her studies, she saw that people tended to interpret robots quite well and without much effort. Intuition helps us, van Otterdijk says.
- However, when the robot’s behaviour contradicts human behaviour, things become difficult. For example, if the robot says "hello" or "have a good time" while brushing someone's hair, people get confused!
An assisting tool, not a replacement
She heard several elderly people say they do not want robots to replace human caregivers. However, after they met her robot, they relaxed, as this did not seem like an option.
Van Otterdijk prefers to see robots as an assisting tool rather than a replacement for human beings. However, she argues that humans will need to think through many ethical questions in the future.
– If robots are to gain some form of intuition, one question we must ask is how they will make their choices. What if a robot brings food to an elderly, who does not feel like eating? Alternatively, what if a person collapses but has expressed that he or she does not want to be helped? Should the robot follow the wishes of that person or the wishes of society?
We should also carefully consider when we want robots to make decisions and when we want humans to make these decisions themselves, she argues.
– Even if we manage to give robots a form of intuition, I do not think that all robots should have it. This depends on the task they are set to perform, Marieke van Otterdijk says.
References:
van Otterdijk, M., Saplacan, D., Baselizadeh, A., Laeng, B., & Torresen, J. (2023, August). To Shake or Not to Shake: Intuitive Reactions of Senior Adults to a Robot Handshake in a Western Culture. In 2023 32nd IEEE International Conference on Robot and Human Interactive Communication (RO-MAN) (pp. 890-896). IEEE. https://doi.org/10.1109/RO-MAN57019.2023.10309608
van Otterdijk, M., Kwak, D., Baselizadeh, A., Lindblom, D. S., & Torresen, J. (2024, August). Age-Old Gesture: Analyzing the Intuitive Responses to Robot Handshakes Among Seniors and Young Adults. In 2024 33rd IEEE International Conference on Robot and Human Interactive Communication (ROMAN) (pp. 1156-1161). IEEE. https://doi.org/10.1109/RO-MAN60168.2024.10731393
van Otterdijk, M., Laeng, B., Lindblom, D. S., & Torresen, J. (2024). The Effect of Expressive Robot Behavior on Users’ Mental Effort: A Pupillometry Study. IEEE Transactions on Cognitive and Developmental Systems. https://doi.org/10.1109/TCDS.2024.3352893