Speaker: Peter Fazekas, AIAS Fellow

The seminar is held in-person, but online attendance is possible via:
https://aarhusuniversity.zoom.us/j/67215252700

Abstract

The recent claim by Google engineer Blake Lemoine that the LaMDA language model is sentient reached headlines all over the world. Categorical refutations followed quickly (and Lemoine got fired), but similar claims in the near future will proliferate. Just like how humans already attribute intentionality and sociality to robots and artificial agents, sentience is also a mental feature that is ascribed automatically on the basis of the impressions that one can get from interacting with these agents. Such ascription is already standard in animal cases and is backed purely by certain similarities in neural and cognitive architecture. Given that state-of-the-art artificial intelligence systems are inspired by some architectural features of human cognitive and neural processes, the question regarding the possibility of sentient machines is not much more hypothetical then the similar question about animal sentience. In this talk I will review what we currently know about human consciousness and how it can help us think about non-human cases.

Short bio

Peter Fazekas is an engineering physicist, a philosopher and a theoretical cognitive scientist with educational background and research experience in computational modelling, artificial intelligence, philosophy of science, philosophy of mind and cognitive neuroscience. He obtained his PhD in philosophy from the University of Edinburgh in 2012

Read about Peter Fazekas's project at AIAS here

What is an AIAS Seminar?

The AIAS Seminar is a session of seminars held by the AIAS fellow or by other speakers proposed by the fellows. In each seminar, one fellow will present and discuss his/her current research and research project, closing off with a question and discussion session.

All seminars are held in English.