We have all been there, pretending to be sick to avoid something that we would rather not do, but what makes pretense unique to humans, and possibly other creatures? Pretense is an ability that humans possess, but also other animals to a certain extent1. For animals, a prime example is playing dead to avoid being eaten. For humans, the desired result could also be to gain attention. A severe example is the Munchausen syndrome, which is a psychological disorder where a person claims to be ill when they are actually lying in order to attract attention and sympathy2. In any case, by pretending, we may use our complex brain and knowledge to imagine people's reactions to our symptoms in order to either gain attention or get out of a situation.

Pretense is an ability that is generally not present among AI or robots, since there is nothing urging them to seek attention or avoid situations. In contrast with humans, robots are made to perform a task, and they do not stop doing so unless it is either ordered to stop or falling apart. Simple robots are not programmed to possess any kind of emotion or capability to feel complex emotions like boredom or irritation and so for them, a task cannot feel ‘tedious’. It can also be noted that robots do not need any kind of motivation to start a task; thus, any form of resistance may not be present as well.

Now that AI is getting more attention due to recent developments such as chatGPT and an AI using human cells learning to play pong3, we found the topic of pretense, and more specifically pretending to be sick, to be intriguing. We wanted to explore what behavioral things set us, humans, apart from AI and robots. But the development of AI using human cells also opens up questions, such as if this type of AI will have consciousness and if the AI will want to avoid situations just as humans do?

Because pretense seems to be a uniquely animal trait, it would therefore be interesting to explore this in the context of robots. How would a robot be perceived as sick, and how could it translate this to humans? The element of pretense is also an interesting and complex type of behavior that could be combined with the perception of being ill. How can a robot pretend like a human? When does a human raise suspicion about the robot lying?

When we started to explore the idea of pretending to be sick, we listed multiple elements of human sickness, such as a high temperature, sneezing and coughing (see list below). A robot probably won’t do any of the listed human indicators, but can you describe errors, warning sounds, or red LEDs as the robot equivalent of being sick? The robot that we have made has some of these human and robotic features. We also added features to the robot that we hoped would increase the sense of compassion that viewers would feel.

We noticed that the human features grabbed the viewers’ attention better. They were, for example, intrigued by a robot that coughs like us, but may not have been convinced that it is actually sick. The features that increased the compassion that the viewers had for the robot were, on the contrary, the robotic elements of being sick. These elements convinced people more that the robot was possibly ill. The viewers felt bad that the robot spelt out ‘error’ in morse code. Maybe the robotic elements created more compassion because they were more realistic to the robot. After the pretending act of the robot ended, people understood that the robot was only acting this way because the robot sensed them to be close by, and thus the robot was perceived as pretending to be ill.