The AI impacts survey records responses from researchers in artificial intelligence and machine learning. When asked, 50% of the researchers who responded thought there was a 5% or higher chance of “future AI advances causing human extinction or similarly permanent and severe disempowerment of the human species”. 50% gave a 10% chance or higher of “human inability to control future advanced AI systems causing human extinction or similarly permanent and severe disempowerment of the human species”.
In another survey of Natural Language Processing researchers:
About a third (36%) of respondents agree that it is plausible that AI could produce catastrophic outcomes in this century, on the level of all-out nuclear war (Q3-4).
I found the links to these results in Eight Things to Know about Large Language Models by Samual Bowman (2023).