The first edition of the “Affordances in the Internet of Things” (AIoT) special session wants to be a container of different contributions coming from distinct research areas related to the world of the Internet of Things (IoT). In particular, the “things” in the “Internet of things” usually contain some knowledge about inputs from the environment and the outputs they need to achieve. For this reason, they “have to know” what will surround them a priori. However, this limits the power of intelligent IoT systems where devices can dynamically learn from their neighbors and change their activity accordingly. In a sense, we would revisit the concept “the whole is not the sum of its parts” in the IoT domain.

The aim of AIoT is to stimulate the debate on AI andKnowledge Representation strategies where aspects like subjectivity, time, roles, and context play some crucial role within the modeling of an IoT system. With AIoT, IoT engineers, computer scientists, psychologists and sociologists are invited to meet each other sharing ideas, problems, and solutions.
The concept of “affordance”, first introduced by James Gibson in the seventies, seems to fit these goals. According to Gibson, objects naturally offer a set of functionalities to the external world depending on the agent that is interacting. In other words, the “meaning” of what happens is the result of a coherent combination of agent-object features. In the IoT, we could see this concept as things that meet each other enabling subsets of their features depending on the context ( i.e., the other objects and the current goals).

Objects have a multitude of behaviors and so they can get to different states or consequences. Some of these consequences obviously coincide with goals, while some others are simply non-relevant with respect to them. However, the latters can enter into the object space of the system and they can be transformed in a new type of input for other objects. For example, the behavior of the window shutters could be “to open” or “to close” according to the weather (in case of rain, the shutters should be closed to maintain the glasses cleaned, or when the sun goes down they could be opened to have more light inside). But, if in the habit there is an air conditioning device, whose scope is to maintain fresh the ambient, the window shutters can receive information on how the conditioner works and "help it” in its functions. In that case, the light that enter may make the ambient warmer, so it could be preferable to have the shutters closed even if it does not rain.
The hope of AIoT is to create an interdisciplinary community around the multi-faceted IoT domain, where different points of view can better push the boundaries towards deeper artificial intelligence systems.