Dear Jonathan Piedra,

The social robot remains a machine, it becomes a (privileged) partner with humans, the existence of interaction with all the elements of its environment; it is part of a dynamic process of interactions that determines how they are disposed towards their human partners and how they are towards the robots with which they interact.

From this there emerge a number of characteristics that must be observed when designing a social robot. In fact, when robots are designed to fit into a social environment, they can no longer be considered as simple autonomous technical objects but as elements that are part of socio-technical systems that they modify (participate fully in events that take place there). This implies that the roboticist takes into account this inclusion of robots (including ethical, legal, social and environmental issues) in the process of designing its functionalities, reliability and performance. It is about taking into consideration the coupling between the object, the people with whom it interacts, the other objects, the environment and the social organization in which it is inserted.

For more details and information about this subject, i suggest you to see links and attached files on topic.

https://searchenterpriseai.techtarget.com/definition/social-robot

https://assets.kpmg/content/dam/kpmg/pdf/2016/06/social-robots.pdf

https://www.emerald.com/insight/content/doi/10.1108/JSM-02-2018-0080/full/html

Article From social brains to social robots: applying neurocognitive...

Chapter Social Robots that Interact with People

Best regards

That's easy, can we label a bot social if it has sentiment about human things.

Dear Mohamed:

Thanks for the comment. Certainly the social robot could not be considered as just any object. But would that give it a different onlogical status? I mean, must the robot actually have social skills (such as dialogue-based communication ability) or just pretend it? A social robot should be openly indeterminate in its functions? or just comply with some rites and norms that we assume as social, for example in the case of caring for the elderly?

Dear Eric:

Thanks for the comment. But I am left wondering what exactly it means (even for a human being) to "have sentiment about human things"?

Well, there are several definitions of "robot" and significant commonalities between them. However, how designers, engineers, policy makers and users of robots make sense of a robot is inherently linked to the functions a robot will perform and the environment within which it will operate. That is not so much to do with the definition of a robot but how one might actually consider building and deploying one. In the end, robots are machines one can think about in the same way as cooking appliances or motor vehicles: the actual size, shape, power, robustness, capacity and specialization are inextricably linked to the problems they are being designed to solve. What we have in addition to these things with robots is an extra layer to the physical and functional profile of a machine: it is the idea of autonomy. Yes, we could think about robot toasters and robot bin lorries.

The way I think about robot is an embodied machine (contrasted with disembodied machine/computational engine), with inputs and outputs to the physical world it occupies. Inputs are given by its array of sensors and the outputs are typically in the form of physical actions. Critically, the robot has a degree of autonomy of action the depends on its programming: logic that determines which actions to select based on inputs, its current operational state and the way it has computed the current state of its operational environment. That logic is key to understanding any robot's autonomy: how much, how predictable, how safe.

What makes social robots distinctive is that they are designed to account for the physical presence of human beings in their operational environments in terms of inputs, logic and outputs. Social robots are often explicitly designed to interact with human beings, rather than to cope with the presence of people as an indirect consequence of sharing their physical space, though my own work anticipates robots entering spaces they will share with people. Human beings are non-deterministic systems, whose range of actions are not trivially reducible to a fixed set of known possible inputs. That presents significant challenges in all respects: how to detect and properly characterise human action, how to select an appropriate response, and how to enact such a response in a safe, timely and accountable manner.

The Rouse definition Dr Lafifi links is helpful in drawing attention to the strength of robots in carrying out routine tasks in accordance with their core programming, and also includes the unpredictability of effects robot may have when encountering situations for which they have not been 'trained' or otherwise programmed. Unfortunately, resolving on predictable situations that permit a high degree of autonomy is very very difficult for social robots. The ethical issues raised by operation in a human environment become a primary concern and extend to the feelings and moral responsibilities of people who will encounter the machines roboticists have created.

The Rouse definition is problematic in a number of ways. Family membership is incompatible with the idea of responsible machine design. Serving as a member of the family suggests moral responsibilities of the robot towards the family, and of other members of the family towards the robot. Prof Joanna Bryson has raised important design concerns about the nature of 'patiency' in this respect:

Article Patiency is not a virtue: the design of intelligent systems ...

I see a significant challenge in Eric Smith's assertion (or assumption) about a social robot as social 'if it has sentiment about human beings'. I think it is inevitable that people will attribute human-like qualities to social robots because anthropomorphism appears to be a robust human trait. It probably makes possible all kinds of sense-making for actually or seemingly autonomous agencies. For a robot to 'have sentiment' would mean it has been programmed with some kind of sentiment analog, to support one or more of its functions. My work has addressed this problem by considering emotional states as coordination devices, drawing on research in animal behaviour, to anticipate benefits for joint human-robot work.

https://www.researchgate.net/project/Framework-for-viable-robot-behaviour-based-inferencing-in-Human-Robot-Collaboration

However, the meaning and value of robot emotional signalling still a huge open question for research in social robotics. For real-world roboticists right now, I'd say it is necessary to take a stance on their design that works to limit attributions of emotion to the machines they intend to create. The world needs neither enthusiastic robot toasters nor indeed grumpy robot bin lorries, however much fun they might at first seem to be.

Dear Jonathan Peidra,

Regarding the capacities, it will be provided with a certain number but with relative limits on hardware and software of robot.

I think for a first phase we refrain from offering the robot functions that are open indefinitely.

Best regards