AI embedded within the social context

Robots can be thought of as artificial intelligences endowed with a body, which act in space in response to the perceptions of their sensors, developing complex relational abilities. Their numbers are growing year on year, whilst their forms and functions are multiplying – in factories, where they have long been present, and increasingly in homes: from lawnmowers and vacuum cleaners to humanoids, from logistics specialists to production operators, and so on. It is highly likely that robots will play a very important role in people’s daily lives by 2050.

But it is even more likely that whatever impact they have on tomorrow’s society will be the result of the choices made today. Design, technology, interfaces, artificial intelligence and everything that will shape the relationship between robots and humans will reflect the values and priorities of those who are designing these machines today and of those who adopt them or plan to introduce them into their social and corporate organisations: these are immense responsibilities. ‘Robotics raises not only questions of technical development, regulatory balance and economic implications, but calls for a civilisational choice on how to organise, over time, the relationship between human beings, machines and society,’ reads the report produced by the IIT in collaboration with Isinnova for Ecosistema Futuro. The potential consequences for work, quality of life, social interaction, the distribution of wealth, the environment itself, and much more, challenge humanity’s ability to imagine the future.

Industrial robots have a long history of operational development, and their use has expanded from the traditional focus on the automotive industry to many other manufacturing sectors. But the major leap forward expected in the coming decades is the advent of humanoid robots or other nature-inspired forms, designed to exist within various ecosystems in the service of humans. These machines lie at the intersection of mechanics and electronics and are attracting the attention of scientists whose research focuses on spatial – also known as physical or embodied – artificial intelligence. People often wonder how the machines set to emerge from this research will relate to human activities, with ambivalent answers ranging from replacing human labour to integrating with people’s roles, particularly for tasks that are dangerous, repetitive or simply of little value.

But it is clear that we must also consider the consequences of the widespread use of robots in terms of the distribution of wealth, the quality of social relations, and how this aligns with foreseeable ecological and demographic changes. In practice, this involves designing the system within a framework of co-evolution, as argued by the authors of the report for Ecosistema Futuro: Giorgio Metta, Scientific Director at IIT; Arash Ajoudani, Principal Investigator of the Human-Robot Interfaces and Interaction Research Unit at IIT; Valentina Malcotti, Head of Content at Isinnova, and Giovanna Giuffrè, Senior Researcher at Isinnova: ‘Without a clear path of co-evolution between work and technology, there is a risk of a gradual erosion of human agency: reduced decision-making and learning abilities, and a diminished sense of purpose in work. Well-designed robotics, on the other hand, can facilitate a reciprocal sharing of the benefits of automation, supporting well-being, ergonomics and the active role of workers.”

The variables to be given priority concern the structures responsible for governing the relationship between people and machines, and the characteristics of the socio-technical system of research and development. Strategic issues include the concentration of control over data and technologies; whether development is geared towards the public or private good; whether the strategy focuses on the interoperability of technologies or on proprietary, closed systems; risk management guided by restrictive regulations or by enabling conditions for responsible experimentation; and shared responsibilities.