Abstract

In the 1990’s, Robotics began to design a new robot aimed at industries (primarily automotive) that worked and interacted with humans outside the cage, thereby replacing traditional _robots_ for some specific duties. This _robot_ is therefore called _co-bot_ (_collaborative_ and _robot)._ Also in the 1990’s, Robotics designed the _social robot_ (for which we propose the neologism _so-bot),_ aimed at assisting humans and keeping them company. The sociality of the _sobots_ lies in their ability to follow the rules of human social life, make decisions independently, and respect the roles assigned to them. Scientific literature usually keeps the terms _collaborative_ and _social robot_ distinct as if they indicated different and separate concepts. We question this separation and affirm that to collaborate (from the Latin _cum-laboro_) means to interact with someone while respecting their nature. Collaboration is that particular form of sociality that relates to work activity. From this it follows that the _cobot_ is essentially social and that _cobots_ and _sobots_ belong to the same category that we call _co-s-bots_ (_collaborative social robots_). In other words, _cobots_ and _sobots_ are two types of _cosbots,_ as the flea and the elephant are two types of animals. The difference between _cobot_ and _sobot_ is given by the development of AI. Both are _potentially_ social, that is, _potentially_ capable of interacting and making decisions independently; but while the _cobot_ is social in potency, the _sobot_ is social ‘in act’. With Aristotelian terminology we can therefore say that the _cobot_ is a _sobot_ in power, while the _sobot_ is a _cobot_ in act. We call this new concept ‘_cobot ontology_’. Such an ontology makes it possible to classify _cobots_ according to the degree of development of AI, just as living beings are classified according to the level of intelligence developed.To teach the _cosbot_ to interact with humans, engineers use some results of neuroscientific research such as mirror neurons and the _embodied Mind_. The use of these models should encourage machine self-learning. Self-learning means autonomy, and autonomy needs strong AI development. It is becoming increasingly clear that autonomy is the condition of the sociality of the _sobot_. The article thus concludes that the relationship between _cobot_ and _sobot_ is the identification of a more general _robot_-_automaton_ (_rabota_-_automatos_) relationship which, in the writer’s opinion, is the essential basis and driving force behind the entire history of Robotics.