We grew up hearing about robots and how, in a few years, they would be an essential part of our lives. When one arrives in the Montefiore Institute, a small robot comes to the visitor and helps him wander around the building until the destination is reached. While doing its task, it will wait for the visitor to follow and will avoid people coming from another direction. For the usual occupant of the Institute, Ragi (see Figure 1.1) has become just another staff member.
Figure 1.1: Ragi
How come this robot, designed only to guide people inside and trained in facial recognition, has managed to integrate the environment and to be trusted by the users? Maybe the cause is linked to the adaptive behaviour of the robot and its interactions with humans, showing a form of intelligence. Does this intelligence and self-awareness of the surroundings require a body?
Ragi is not the only robot with a body or human features, such as its big eyes. The idea is in fact quite old. For example, in 1770, you could have played chess against a Mechanical Turk who had all the physical traits of a real human thus increasing the level of interaction.
In our quest to build new robots, we are trying to create intelligent machines equipped with a body, embracing the concept of “Embodied Intelligence“. There are those who argue that they will be an undeniable advantage to society, and there are also those who fear that they will become more intelligent than us and that they will control humanity. Embodied Intelligence refers to “a computational approach to the design and understanding of intelligent behaviour in embodied and situated agents through the consideration of the strict coupling between the agent and its environment (situatedness), mediated by the constraints of the agent’s own body, perceptual and motor system, and brain (embodiment)”. 
Embodied Intelligence promotes the view that the body, the information we take from the physical world, and our individual experiences with the environment all play a central role in intelligence and learning. This view raises important questions related to the limits of artificial intelligence. To what extent can we create a robot like a human being?
The series Westworld is a great example that can make us think about this, raising several ethical questions. Imagine that, just like in Westworld, we were able to faithfully reproduce a robot in human form: physical appearance, emotions, and feelings. Shouldn’t they be considered intelligent beings, like us, with the right to life and freedom? Throughout the series this topic is widely discussed, and on one side we have the perspective of Dolores, a robot trapped for years in an entertainment park for humans, who is beginning to become aware of herself and her condition. On the other side we have Dr. Robert Ford, the designer of Dolores and many other robots like her, who believes in the potential of his creation and is willing to go to any extent to protect it, even if it means harming humanity. In the middle of this duo, we have characters who are against the increasing development of robots and defend limits to their evolution, while others are curious about their potential and how far they can go.
The truth is that, even if we are talking about a series, this debate also takes place in the real world, although on a smaller scale, and the question can be raised to which extent this series is a fiction or a foreseeing of a near future. Embodied intelligence is a reality that keeps growing, raising a lot of important questions.
Why did the idea of Embodied Intelligence emerged? Which problems did this new approach solve? How can it be a solution to classical Artificial Intelligence? Does a robot need to feel emotions? What does it imply for intelligence to be embodied? What are the differences between an intelligence that works through interaction and one that does through data? What advances have already been made in this area? What improvements can be made, for example to increase their robustness? What is the importance of the likeness of the robot in the interaction with humans? Does the robot only have to interact with being or could it interact with its kind?
The present blog explores all those questions. First, it reviews the origins and motivations behind Embodied Intelligence. Then the blog covers the heritage of Rodney Brooks, an Australian roboticist, who is considered as the father of the current vision of Embodied Intelligence. Finally, the requirement of embodiment and human-robots interaction are analysed, leading to the discussion of Embodied Intelligence’s future and the perspective that it offers.