New PAIR Publication on AI Risk
PAIR research fellow Dr. Leonard Dung has published a new paper titled ‘How to deal with risks of AI suffering‘ (https://www.tandfonline.com/doi/abs/10.1080/0020174X.2023.2238287?journalCode=sinq20). In this paper, published in the philosophy journal Inquiry, Dung argues that humans might soon create AI systems which are sentient and able to suffer. Since such systems could be easily copied, this creates a significant moral risk of massive suffering. In response, Dung proposes, discusses and evaluates different strategies to reduce this risk.