top of page

Thu, 11 Oct

|

Room 234

Matthew Crosby (Imperial) 4I Consciousness

From phenomenological axioms to model-building systems, a framework for researching AI consciousness

Time & Location

11 Oct 2018, 16:00 – 18:00

Room 234, Senate House, London WC1B, UK

Abstract

Integrated Information Theory (IIT) is an influential theory of consciousness with a unique methodology. It starts from a set of phenomenological axioms, meant to be self-evident truths about all possible experiences, and, from these, derives a quantitative measure of consciousness based on the cause-effect properties of physical systems. While IIT has many advocates, there are numerous criticisms of the theory. IIT is reliant on the accuracy of its initial axioms, and the translation from phenomenology to cause-effect properties of matter. As the explanandum is so far removed from the explanans, it is hard to verify the translation process. The aim of this paper is to build on the successes of IIT’s unique methodology, whilst alleviating the translation problem by setting the explanans as representational properties of model-building systems, an intermediate level between phenomenology and the cause-effect properties of matter. The axioms are also changed to pertain to ‘comprehensible consciousness’ as opposed to all possible conscious experience. The result is no longer a quantitative theory of consciousness in general, but a theory, 4I consciousness, that can determine which systems are consciousness candidates and can be used to guide a research agenda for understanding consciousness in the future of artificial intelligence.

Share this event

bottom of page