In our everyday life, we use simple tools to make our life easier, safer, or more independent. One example of a simple tool is the screwdriver. It is designed to pursue a specific goal, to turn screws, it cannot accomplish that goal without us, and it is unlikely to surprise us. Robots, on the contrary, are not simple tools. They can be designed to pursue multiple goals, often without us, and they have a more significant opportunity to surprise us. Robots with autonomous behaviors can do much more, but they have their own set of drawbacks. For instance, one major shortcoming is the possibility that humans working with a robot can fall “out of the loop,” completely delegating a task to the robot and losing awareness of that task and the robot. Even if a robot is reliable, its failure to communicate relevant information about itself and its actions with us will inhibit our ability to work interdependently with it.
We must understand robots’ intentions to provide a smooth, comfortable, and especially safe interaction or co-existence in such situations. Therefore, a robot must transparently communicate its intended actions to reduce uncertainty.
Robots need the fusion of legibility, explainability, and predictability to reach transparency.
Accordingly, interacting with a transparent robot, we, humans, should be able to answer three questions:
What does the robot do?
Why is this robot doing this?
What will the robot do next?
Is it possible to answer the above questions for this robot?
Such questions open a major inquiry: how will roboticists create such a transparent robot? The answer lies in the Human-Human Interaction field and how we interact with each other.
A large part of human communication takes place non-verbally. Non-verbal behaviors such as gaze, facial expressions, and gestures during Human-Human Interaction carry essential knowledge regarding intentions, action goals, and focus of attention. Analyzing the basic principles of human communication through non-verbal signals and integrating them into robots can be an effective solution to tackle the problem of transparency due to the familiarity these behaviors have in humans.
This issue is as arduous as compelling, and successful results could have a phenomenal impact on integrating robots into our society.