Free-Energy Minimising Agents and Beneficial A.I.: Ambient Smart Environments, Allostasis, and Metacognitive Control
Year: 2023
Type: preprint
Abstract: Emerging forms of technology are beginning to change the ways that humans engage with and rely on their material environments. In this paper we focus on material interior environments that are infused with smart technology, which we call ambient smart environments (ASEs). ASEs can track their users, gather data, and intervene in shaping everyday opportunities for action. We argue that a traditional approach to theorising technology, using the language of affordances, focusing on what technology allows us to do, fails to fully capture the potential role ASEs could come to play in people’s lives. This is because a primary design characteristic of ASEs is the drive to dissolvetraditional user interfaces, and to have the technology recede into the background. We argue for a view of ASEs as a meta-affordance, actively shaping their user’s field of affordances. To develop our account, we draw on a wide range of theoretical cognitive science, including The Skilled Intentionality Framework and the Free Energy Principle. We argue that the true potential of ASEs lies in their capacity to intervene on an agent’s field of affordances, thus playing an integral role in the agent’s capacity to minimise free energy effectivity.
Authors: Ben White, Mark Miller
Institutions: University of Sussex, Monash University
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: gold
Funder: Leverhulme Trust
OpenAlex ID: W4317606103