AI Agency and Wellbeing Workshop
The workshop explores the following questions:
- Can a large language model be a cognitive and linguistic agent (in the way that humans are agents)?
- Can there be such a thing as AI well-being?
- If the answers are yes, what are the implications?
Date: November 22, 2023
Title: Varieties of Moral Agency and Risks of Digital Dystopia
Speaker: Dr Adam Bradley, Lingnan University
Abstract:
We argue that AIs will likely soon possess a form of moral agency—interest-conferring agency—that bestows them with distinctive moral interests (rights, welfare). This fact has important ethical consequences because the emergence of agency-conferred interests in AIs will bring with it the potential for dystopian moral catastrophes. We identify and describe three in particular. First, there is a threat of artificial absurdity, a condition in which AIs have self-conceptions that are disconnected from reality in a way that detracts significance from their lives. Second, there are oppressive scenarios in which AI’s have rights that are grounded in their agency and violated on a large scale. Third, there are distributive disasters in which the distribution of different forms of moral agency results in an unjust society. We contend that these risks deserve to be considered alongside more familiar existential and suffering risks, and that urgent steps must be taken to mitigate them in order to avoid a digital dystopia.