Artificial Sociality: simulating the social mind
Self-organisation is the phenomenon, observed in many systems, that the whole is more than the sum of its parts. The whole self-organises as the parts interact, without anyone being in control. All living systems self-organise, from cells to societies. This leads to ‘elephant paths’.
Social simulation uses models that self-organise: Agent-based models. ‘Agents’ are the elements, such as people, that are given simple motivations and interactions in a virtual world; the model is run across time; and the resulting system-level patterns can be visualised and analysed. The dynamics of socio-ecological and socio-technical systems can be better understood with agent-based models. Agent-based modelling is becoming popular for policy support. It can reveal the
‘elephant paths’ by which society self-organises.
Finally, Gert Jan zooms in to the social side of these agent-based models. What social motives should the agents have? This is the issue of artificial sociality, the social counterpart of artificial intelligence. Gert Jan will discuss the state of the art, present his meta-model GRASP (Groups – Rituals – Affiliation – Significance – Power), and show some current simulations.
The conclusion is that we humans still are quite naïve about the unintended, self-organized collective consequences of our behaviours; and we should urgently change this.