If you imagine an agent as a blackbox, you can first focus on the inputs it can receive and the outputs it can produce. In a second step, you then peek under the hood to understand what happens inside that box, how the inputs are connected to the outputs. Let us begin with the agent’s relations to the outside: What goes in? What comes out?
In order to make sense of the world, an agent first has to sense the world; it needs some sensory equipment to ‘read’ the context. We humans rely on our proverbial five senses: we see, hear, smell, taste, and touch. That’s it. Many natural agents employ similar means, but others evolved the capacity to sense for example humidity, acidity, salinity, gravity, electrical current, magnetic fields, or chemical traces. It is important to recognise that much more is possible than our anthropocentric emphasis on vision and on abstract information leads us to believe. Yet for most artificial agents, their ‘view of the world’ is data- and information-centric. This reduced view of the outside world enhances the efficiency of their inner workings. The jury is still out whether our self-made artefacts are thus sufficiently equipped to make sense of the world.
To effect a change in the environment, an agents needs to ‘do something’. It must be able to act, have some tools to instigate a change. From our anthropocentric venture point, we perceive of action as some force or motion that we can easily observe, that is strong enough and neither too fast nor too slow for our senses to register. So we tend to imagine a muscle or a robotic arm, to think rather narrowly about ‘something mechanical’. But physics includes everything electrical and magnetic; and there’s the entire realm of chemistry as well. The possible action modes are as diverse as the sensing modes indicated above. And the individual agent is not necessarily alone. It can extend its range and reach through indirect action, i.e., by triggering another agent. Consider for example social insects, such as ants communicating with scents, or honey bees and their waggle dance. Amongst the natural agents, herds, flocks, and swarms are ‘teams’ made up of the same species, whereas symbiosis describes a mutually beneficial cooperation across species boundaries. The same baseline interactions exist between humans and artificial agents, only that they reach far higher levels of complexity and abstraction. Yet ultimately, the available tools, no matter how mighty, constrain any agent’s option space.
In short, the world goes in, and an action comes out. However, the perception of the world is filtered through the capacity of the agent’s sensory equipment, and thus considerably reduced. Similarly, the possible action is limited by the capacity of the manipulators available to the agent, whether directly or indirectly.
This is the eleventh in a series of posts on the agency and how it matters to innovation.
understanding innovation 